Download gSOAP 2.7.9 User Guide

gSOAP 2.7.9 User Guide
Robert van Engelen
Florida State University
and Genivia, Inc.
[email protected] & [email protected]
June 26, 2007
Contents
1 Introduction
8
1.1
Getting Started . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8
1.2
Your First Web Service Client Application . . . . . . . . . . . . . . . . . . . . . .
9
1.3
Your First Web Service in CGI . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10
1.4
Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11
2 Notational Conventions
13
3 Differences Between gSOAP Versions 2.4 (and Earlier) and 2.5
14
4 Differences Between gSOAP Versions 2.1 (and Earlier) and 2.2
14
5 Differences Between gSOAP Versions 1.X and 2.X
14
6 Interoperability
17
7 Quick User Guide
18
7.1
How to Use the gSOAP Stub and Skeleton Compiler to Build SOAP Clients . . .
18
7.1.1
Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
19
7.1.2
Namespace Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . .
23
7.1.3
Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
25
7.1.4
How to Generate C++ Client Proxy Classes . . . . . . . . . . . . . . . . .
26
7.1.5
XSD Type Encoding Considerations . . . . . . . . . . . . . . . . . . . . .
27
7.1.6
Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
28
7.1.7
How to Change the Response Element Name . . . . . . . . . . . . . . . .
29
7.1.8
Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
29
1
7.1.9
7.2
How to Specify Multiple Output Parameters . . . . . . . . . . . . . . . .
30
7.1.10 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
30
7.1.11 How to Specify Output Parameters With struct/class Compound Data
Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
31
7.1.12 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
32
7.1.13 How to Specify Anonymous Parameter Names . . . . . . . . . . . . . . .
34
7.1.14 How to Specify a Method with No Input Parameters . . . . . . . . . . . .
35
7.1.15 How to Specify a Method with No Output Parameters . . . . . . . . . . .
36
How to Use the gSOAP Stub and Skeleton Compiler to Build SOAP Web Services 36
7.2.1
Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
36
7.2.2
MSVC++ Builds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
39
7.2.3
How to Create a Stand-Alone gSOAP Service . . . . . . . . . . . . . . . .
39
7.2.4
How to Create a Multi-Threaded Stand-Alone Service . . . . . . . . . . .
41
7.2.5
How to Pass Application Data to Service Methods . . . . . . . . . . . . .
47
7.2.6
Some Web Service Implementation Issues . . . . . . . . . . . . . . . . . .
47
7.2.7
How to Generate C++ Server Object Classes . . . . . . . . . . . . . . . .
48
7.2.8
How to Generate WSDL Service Descriptions . . . . . . . . . . . . . . . .
49
7.2.9
Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
50
7.2.10 How to Parse and Import WSDL Service Descriptions to Develop Clients
and Servers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
52
7.2.11 The typemap.dat File . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
54
7.2.12 How to Use Client Functionalities Within a Service . . . . . . . . . . . . .
54
7.3
How to Use gSOAP for Asynchronous One-Way Message Passing . . . . . . . . .
57
7.4
One-Way Message Passing over HTTP . . . . . . . . . . . . . . . . . . . . . . . .
58
7.5
How to Use the SOAP Serializers and Deserializers to Save and Load Application
Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
58
7.5.1
Serializing a Data Type . . . . . . . . . . . . . . . . . . . . . . . . . . . .
59
7.5.2
Deserializing a Data Type . . . . . . . . . . . . . . . . . . . . . . . . . . .
64
7.5.3
Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
65
7.5.4
Serializing and Deserializing Class Instances to Streams . . . . . . . . . .
69
7.5.5
How to Specify Default Values for Omitted Data . . . . . . . . . . . . . .
70
8 Using the gSOAP Stub and Skeleton Compiler
72
8.1
Compiler Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
73
8.2
SOAP 1.1 Versus SOAP 1.2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
74
8.3
The soapdefs.h Header File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
75
8.4
How to Build Modules and Libraries with the gSOAP #module Directive . . . .
75
8.5
How to use the gSOAP #import Directive . . . . . . . . . . . . . . . . . . . . . .
76
8.6
How to Use #include and #define Directives . . . . . . . . . . . . . . . . . . . .
77
2
8.7
Compiling a gSOAP Client . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
77
8.8
Compiling a gSOAP Web Service . . . . . . . . . . . . . . . . . . . . . . . . . . .
78
8.9
Using gSOAP for Creating Web Services and Clients in Pure C . . . . . . . . . .
79
8.10 Limitations of gSOAP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
79
8.11 Compile Time Flags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
81
8.12 Run Time Flags
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
81
8.13 Memory Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
83
8.13.1 Memory Management Policies . . . . . . . . . . . . . . . . . . . . . . . . .
84
8.13.2 Intra-Class Memory Management . . . . . . . . . . . . . . . . . . . . . . .
86
8.14 Debugging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
88
8.15 Required Libraries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
89
9 The gSOAP Remote Method Specification Format
89
9.1
Remote Method Parameter Passing . . . . . . . . . . . . . . . . . . . . . . . . . .
90
9.2
Error Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
92
9.3
C/C++ Identifier Name to XML Name Translations . . . . . . . . . . . . . . . .
95
9.4
Namespace Mapping Table . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
96
10 gSOAP Serialization and Deserialization Rules
98
10.1 SOAP RPC Encoding Versus Document/Literal and xsi:type Info . . . . . . . . .
98
10.2 Primitive Type Encoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
99
10.3 How to Encode and Decode Primitive Types as XSD Types . . . . . . . . . . . .
99
10.3.1 How to Use Multiple C/C++ Types for a Single Primitive XSD Type . . 106
10.3.2 How to use Wrapper Classes to Specify Polymorphic Primitive Types . . 106
10.3.3 XSD Schema Type Decoding Rules . . . . . . . . . . . . . . . . . . . . . . 108
10.3.4 Multi-Reference Strings . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
10.3.5 “Smart String” Mixed-Content Decoding . . . . . . . . . . . . . . . . . . 111
10.3.6 STL Strings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
10.3.7 Changing the Encoding Precision of float and double Types . . . . . . . 112
10.3.8 INF, -INF, and NaN Values of float and double Types . . . . . . . . . . 113
10.4 Enumeration Serialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
10.4.1 Serialization of Symbolic Enumeration Constants . . . . . . . . . . . . . . 113
10.4.2 Encoding of Enumeration Constants . . . . . . . . . . . . . . . . . . . . . 114
10.4.3 Initialized Enumeration Constants . . . . . . . . . . . . . . . . . . . . . . 115
10.4.4 How to “Reuse” Symbolic Enumeration Constants . . . . . . . . . . . . . 115
10.4.5 Boolean Enumeration Serialization for C . . . . . . . . . . . . . . . . . . . 116
10.4.6 Bitmask Enumeration Serialization . . . . . . . . . . . . . . . . . . . . . . 116
10.5 Struct Serialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
10.6 Class Instance Serialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
3
10.6.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
10.6.2 Initialized static const Fields . . . . . . . . . . . . . . . . . . . . . . . . 119
10.6.3 Class Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
10.6.4 Getter and Setter Methods . . . . . . . . . . . . . . . . . . . . . . . . . . 120
10.6.5 Streaming XML with Getter and Setter Methods . . . . . . . . . . . . . . 121
10.6.6 Polymorphism, Derived Classes, and Dynamic Binding . . . . . . . . . . . 122
10.6.7 XML Attributes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
10.6.8 QName Attributes and Elements . . . . . . . . . . . . . . . . . . . . . . . 127
10.7 Union Serialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
10.8 Serializing Pointer Types
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
10.8.1 Multi-Referenced Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
10.8.2 NULL Pointers and Nil Elements . . . . . . . . . . . . . . . . . . . . . . . 130
10.9 Void Pointers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
10.10Fixed-Size Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
10.11Dynamic Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
10.11.1 SOAP Array Bounds Limits . . . . . . . . . . . . . . . . . . . . . . . . . . 133
10.11.2 One-Dimensional Dynamic Arrays . . . . . . . . . . . . . . . . . . . . . . 133
10.11.3 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
10.11.4 One-Dimensional Dynamic Arrays With Non-Zero Offset
. . . . . . . . . 136
10.11.5 Nested One-Dimensional Dynamic Arrays . . . . . . . . . . . . . . . . . . 137
10.11.6 Multi-Dimensional Dynamic Arrays . . . . . . . . . . . . . . . . . . . . . 138
10.11.7 Encoding XML Generics Containing Dynamic Arrays . . . . . . . . . . . 139
10.11.8 STL Containers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
10.11.9 Polymorphic Dynamic Arrays and Lists . . . . . . . . . . . . . . . . . . . 143
10.11.10How to Change the Tag Names of the Elements of a SOAP Array or List 143
10.12Base64Binary XML Schema Type Encoding . . . . . . . . . . . . . . . . . . . . . 144
10.13hexBinary XML Schema Type Encoding . . . . . . . . . . . . . . . . . . . . . . . 146
10.14Literal XML Encoding Style . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
10.14.1 Serializing and Deserializing Mixed Content XML With Strings . . . . . . 148
11 SOAP Fault Processing
150
12 SOAP Header Processing
152
13 MIME Attachments
154
13.1 Sending a Collection of MIME Attachments (SwA) . . . . . . . . . . . . . . . . . 155
13.2 Retrieving a Collection of MIME Attachments (SwA) . . . . . . . . . . . . . . . 157
4
14 DIME Attachments
158
14.1 Sending a Collection of DIME Attachments . . . . . . . . . . . . . . . . . . . . . 158
14.2 Retrieving a Collection of DIME Attachments . . . . . . . . . . . . . . . . . . . . 159
14.3 Serializing Binary Data in DIME . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
14.4 Streaming DIME . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
14.5 Streaming Chunked DIME . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
14.6 WSDL Bindings for DIME Attachments . . . . . . . . . . . . . . . . . . . . . . . 166
15 MTOM Attachments
166
15.1 Generating MultipartRelated MIME Attachment Bindings in WSDL . . . . . . . 168
15.2 Sending and Receiving MTOM Attachments . . . . . . . . . . . . . . . . . . . . . 168
15.3 Streaming MTOM/MIME . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
15.4 Redirecting Inbound MTOM/MIME Streams Based on SOAP Body Content . . 174
15.5 Streaming Chunked MTOM/MIME . . . . . . . . . . . . . . . . . . . . . . . . . 175
16 XML Validation
176
16.1 Occurrence Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
16.1.1 Elements with minOccurs and maxOccurs Restrictions . . . . . . . . . . . 176
16.1.2 Required and Prohibited Attributes . . . . . . . . . . . . . . . . . . . . . 177
16.1.3 Data Length Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
16.2 Other Constraints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
17 SOAP-over-UDP
179
17.1 Using WS-Addressing with SOAP-over-UDP . . . . . . . . . . . . . . . . . . . . 180
17.2 Client-side One-way Unicast . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
17.3 Client-side One-way Multicast . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 181
17.4 Client-side Request-Response Unicast . . . . . . . . . . . . . . . . . . . . . . . . 181
17.5 Client-side Request-Response Multicast . . . . . . . . . . . . . . . . . . . . . . . 182
17.6 SOAP-over-UDP Server . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
18 Advanced Features
185
18.1 Internationalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
18.2 Customizing the WSDL and Namespace Mapping Table File Contents With
gSOAP Directives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
18.2.1 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
18.3 Transient Data Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
18.4 Volatile Data Types . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
18.5 How to Declare User-Defined Serializers and Deserializers . . . . . . . . . . . . . 195
18.6 How to Serialize Data Without Generating XSD Type Attributes . . . . . . . . . 196
18.7 Function Callbacks for Customized I/O and HTTP Handling . . . . . . . . . . . 196
5
18.8 HTTP 1.0 and 1.1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
18.9 HTTP 307 Temporary Redirect Support . . . . . . . . . . . . . . . . . . . . . . . 202
18.10HTTP GET Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
18.11HTTP Keep-Alive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
18.12HTTP Chunked Transfer Encoding . . . . . . . . . . . . . . . . . . . . . . . . . . 206
18.13HTTP Buffered Sends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
18.14HTTP Authentication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
18.15HTTP Proxy Authentication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
18.16Speed Improvement Tips . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
18.17Timeout Management for Non-Blocking Operations . . . . . . . . . . . . . . . . . 208
18.18Socket Options and Flags . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
18.19Secure SOAP Web Services with HTTPS/SSL . . . . . . . . . . . . . . . . . . . . 209
18.20Secure SOAP Clients with HTTPS/SSL . . . . . . . . . . . . . . . . . . . . . . . 214
18.21SSL Authentication Callback . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
18.22SSL Certificates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
18.23SSL Hardware Acceleration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
18.24SSL on Windows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
18.25Zlib Compression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
18.26Client-Side Cookie Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
18.27Server-Side Cookie Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 220
18.28Connecting Clients Through Proxy Servers . . . . . . . . . . . . . . . . . . . . . 223
18.29FastCGI Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
18.30How to Create gSOAP Applications With a Small Memory Footprint . . . . . . . 223
18.31How to Eliminate BSD Socket Library Linkage . . . . . . . . . . . . . . . . . . . 224
18.32How to Combine Multiple Client and Server Implementations into one Executable 225
18.33How to Build a Client or Server in a C++ Code Namespace . . . . . . . . . . . . 226
18.34How to Create Client/Server Libraries . . . . . . . . . . . . . . . . . . . . . . . . 227
18.34.1 C++ Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 228
18.34.2 C Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
18.35How to Create DLLs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
18.35.1 Create the Base stdsoap2.dll . . . . . . . . . . . . . . . . . . . . . . . . . 233
18.35.2 Creating Client and Server DLLs . . . . . . . . . . . . . . . . . . . . . . . 233
18.36gSOAP Plug-ins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234
18.36.1 The Message Logging and Statistics Plug-in . . . . . . . . . . . . . . . . . 236
18.36.2 The HTTP GET Plug-in . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
18.36.3 The HTTP MD5 Plug-in . . . . . . . . . . . . . . . . . . . . . . . . . . . 238
18.36.4 The HTTP Digest Authentication Plug-in . . . . . . . . . . . . . . . . . . 239
18.36.5 The WS-Addressing Plug-in . . . . . . . . . . . . . . . . . . . . . . . . . . 240
18.36.6 The WS-Security Plug-in . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
6
Copyright (C) 2000-2006 Robert A. van Engelen, Genivia, Inc., All Rights Reserved.
7
1
Introduction
The gSOAP tools provide a SOAP/XML-to-C/C++ language binding to ease the development
of SOAP/XML Web services and client application in C and C++. Most toolkits for C++ Web
services adopt a SOAP-centric view and offer APIs that require the use of class libraries for SOAPspecific data structures. This often forces a user to adapt the application logic to these libraries. In
contrast, gSOAP provides a C/C++ transparent SOAP API through the use of compiler technology
that hides irrelevant SOAP-specific details from the user. The gSOAP stub and skeleton compiler
automatically maps native and user-defined C and C++ data types to semantically equivalent XML
data types and vice-versa. As a result, full SOAP interoperability is achieved with a simple API
relieving the user from the burden of SOAP details, thus enabling him or her to concentrate on the
application-essential logic.
The gSOAP compiler enables the integration of (legacy) C/C++ and Fortran codes (through a
Fortran to C interface), embedded systems, and real-time software in SOAP applications that share
computational resources and information with other SOAP applications, possibly across different
platforms, language environments, and disparate organizations located behind firewalls.
1.1
Getting Started
To start building Web services applications with gSOAP, you need:
• The gSOAP package from http://sourceforge.net/projects/gsoap2
• A C or C++ compiler.
• You may want to install OpenSSL and the Zlib libraries to enable SSL (HTTPS) and compression. These libraries are available for most platforms and are often already installed.
gSOAP is self-contained, so there is no need to download any third-party software (unless you want
to use OpenSSL and the library is not already installed).
Although gSOAP is available in binary format for several platforms, the code generated by the
gSOAP stub and skeleton compiler and the gSOAP runtime codes are equivalent. This means that
the generated codes can be transferred to other platforms and compiled.
The gSOAP packages available from SourceForge include pre-build tools:
• The wsdl2h WSDL/schema parser tool.
• The soapcpp2 stub/skeleton compiler.
Win32 versions of these two are included in the Win32 gSOAP package only. If you don’t have the
binaries or if you want to rebuild them, you need
• A C++ compiler to build wsdl2h.
• A C compiler and Bison or Yacc to build soapcpp2.
8
• A C compiler and Flex or Lex to build soapcpp2.
Bison and Flex are preferred.
The tools are used to generate code that is linked with the gSOAP engine soapcpp2.c (C version)
or soapcpp2.cpp (C++ version) and your application code. The engine is also available as a library
libgsoap.a and libgsoap++.a with separate versions that support SSL. See the README.txt instructions
on how to build these libraries with the platform-independent gSOAP package’s autoconf and
automake.
The gSOAP packages contain numerous examples in the samples directory. Run make to build the
example applications. The examples are also meant to demonstrate different features of gSOAP.
The simplest examples are the one-liners in samples/oneliners. Indeed, you can write a one-line
Web service with CGI!. A streaming DIME attachment server and client application demonstrate
efficient file exchanges in samples/dime. An SSL-secure Web server application demonstrates the
generation of dynamic content for Web browsing and Web services functionality at the same time,
see samples/webservice. And much more.
1.2
Your First Web Service Client Application
The gSOAP tools minimize application adaptation efforts for building Web Services. The gSOAP
wsdl2h tool imports one or more WSDLs and XML schemas to generate a header file with the Web
service operations and the C/C++ data types used by the services. The gSOAP soapcpp2 compiler
takes the header file and generates XML serializers for the data types (soapH.h and soapC.cpp), the
client-side stubs (soapClient.cpp), and server-side skeletons (soapServer.cpp).
The gSOAP soapcpp2 compiler can also generate WSDL definitions for implementing a service from
scratch, i.e. without defining a WSDL first. This ”closes the circle” in that it enables Web services
development from WSDL or directly from a set op C/C++ operations in a header file without the
need for users to analyze Web service details.
You only need to follow a few steps to execute the tools from the command line or Makefile (see also
MSVC++ project examples in the samples directory with tool integration in the MSVC++ IDE).
For example, to generate code for the XMethods service listing Web service, we run the wsdl2h tool
from the command line on the URL of the WSDL and use option -o to specify the output file:
$ wsdl2h -o XMethodsQuery.h http://www.xmethods.net/wsdl/query.wsdl
This generates the XMethodsQuery.h header file with Web service operations and the data types that
the service uses. This header file is to be processed with soapcpp2 to generate the stub and/or
skeleton code. The XMethodsQuery.h file includes all documentation, so you can use Doxygen (http:
//www.doxygen.org) to automatically generate the documentation pages for your development.
In this example we are developing a C++ API for the XMethods service. By default, gSOAP
assumes you will use C++ with STL. To build without STL, use option -s:
$ wsdl2h -s -o XMethodsQuery.h http://www.xmethods.net/wsdl/query.wsdl
To build a pure C application, use option -c:
9
$ wsdl2h -c -o XMethodsQuery.h http://www.xmethods.net/wsdl/query.wsdl
We have not yet generated the stubs for the C/C++ API. To do so, run the soapcpp2 compiler:
$ soapcpp2 -C -Iimport XMethodsQuery.h
Where option -C indicates client-side only files (soapcpp2 generates both client and server stubs and
skeletons by default). Option -I is needed to import the stlvector.h file to support STL vectors.
Suppose we develop a C++ client for the XMethods service. In this case we use the generated
soapXMethodsQuerySoapProxy class and XMethodsQuerySoap.nsmap XML namespace mapping table to
access the Web service. The soapXMethodsQuerySoapProxy class is a proxy to invoke the service:
#include ”soapXMethodsQuerySoapProxy.h”
#include ”XMethodsQuerySoap.nsmap”
main()
{
XMethodsQuerySoap service;
ns3 getAllServiceNamesResponse response;
// get all service names from the XMethods database:
if (service.ns3 getAllServiceNames(response) == SOAP OK)
std::cout << ”The first XMethods service is: ” << (*response. Result-> ptr[0]->name) <<
std::endl;
else
soap print fault(service.soap, stderr);
}
The response data structure is defined in XMethodsQuery.h, and contains a SOAP encoded array
( ptr[n]) of pointers to ID-Name pairs ( ptr[n]->id and ptr[n]->name). (Note: you may want to
add NULL checks before dereferencing the pointers.) To complete the build, compile and link the
generated soapC.cpp, soapClient.cpp, and the run-time gSOAP engine stdsoap2.cpp with your code.
1.3
Your First Web Service in CGI
Developing a service application is easy too.
Suppose we implement a CGI-based service that returns the time in GMT. The Common Gateway
Interface (CGI) is a simple mechanism to publish services on your Web site, but it is certainly not
the most efficient way. You can also develop high-performance stand-alone gSOAP services with
built-in HTTP/S stacks or you can use Apache mod gsoap and IIS (see the extras directory).
Our currentTime service only has an output parameter, which is the current time:
// File: currentTime.h
//gsoap ns service name: currentTime
//gsoap ns service namespace: urn:currentTime
//gsoap ns service location: http://www.yourdomain.com/currentTime.cgi
int ns currentTime(time t& response);
Note that we must associate an XML namespace with a service. The gSOAP tools use a special
convention for identifier names that are part of a namespace: a namespace prefix (ns in this case)
10
followed by a double underscore . This convention is used to resolve namespaces and to avoid
name clashes. The ns namespace prefix is bound to the urn:currentTime namespace name with the
//gsoap directive. The //gsoap directives are used to set the properties of the service, in this case
the name, namespace, and location endpoint.
The service implementation for CGI is
// File: currentTime.cpp
main()
{
// create soap context and serve one CGI-based request:
soap serve(soap new());
}
int ns currentTime(struct soap *soap, time t& response)
{
response = time(0);
return SOAP OK;
}
Note that we pass the soap struct with the gSOAP context information to the service routine,
which can be handy to determine properties of the connection and to dynamically allocate data
with soap malloc(soap, num bytes) that will be automatically deleted when the service is finished. We
run the soapcpp2 compiler on the header file to generate the server-side code:
$ soapcpp2 -S currentTime.h
and then compile the CGI binary:
$ c++ -o currentTime.cgi currentTime.cpp soapC.cpp soapServer.cpp stdsoap2.cpp
To activate the service, copy the currentTime.cgi binary to your bin-cgi directory with the proper file
permissions.
The soapcpp2 compiler generated the WSDL definitions currentTime.wsdl. You can use the WSDL to
advertize your service. You don’t need to use this WSDL to develop a gSOAP client. You can use
the currentTime.h file with the soapcpp2 -C command to generate client-side code.
When you contribute a Web service with interesting capabilities, you can contact www.XMethods.com
to publish your service and see it with the client application for the XMethods service listing you
developed in the previous section.
1.4
Features
The highlights of gSOAP are:
• Unique interoperability features: the gSOAP compiler generates SOAP marshalling routines
that (de)serialize native and user-defined C and C++ data structures.
• gSOAP supports WSDL 1.1, SOAP 1.1, SOAP 1.2, SOAP RPC encoding style, and document/literal style.
11
• gSOAP is one of the few SOAP toolkits that support the full range of SOAP 1.1 RPC encoding
features including sparse multi-dimensional arrays and polymorphic types. For example, a
remote method with a base class parameter may accept derived class instances from a client.
Derived class instances keep their identity through dynamic binding.
• gSOAP supports MIME (SwA), DIME, and MTOM attachments and has streaming capabilities to direct the data stream to/from resources.
• gSOAP is the only toolkit that supports streaming DIME attachment transfers, which allows you to exchange binary data of practically unlimited size in the fastest possible way
(streaming) while ensuring the usefulness of XML interoperability.
• gSOAP supports SOAP-over-UDP.
• gSOAP supports IPv4 and IPv6.
• gSOAP supports Zlib deflate and gzip compression (for HTTP, TCP/IP, and XML file storage).
• gSOAP supports SSL (HTTPS).
• gSOAP supports HTTP/1.0, HTTP/1.1 keep-alive, chunking, basic authentication, and digest
authentication using a plugin.
• gSOAP supports SOAP one-way messaging.
• The schema-specific XML pull parser is fast and efficient and does not require intermediate
data storage for demarshalling to save space and time.
• The gSOAP soapcpp2 compiler includes a WSDL generator for convenient Web Service publishing.
• gSOAP includes a WSDL parser wsld2h (WSDL converter to gSOAP header files) for automated client and server development.
• Generates source code for stand-alone Web Services and client applications.
• Ideal for small devices such as Palm OS, Symbian, Pocket PC, because the memory footprint
is small.
• Ideal for building web services that are compute-intensive and are therefore best written in
C and C++.
• Platform independent: Windows, Unix, Linux, Mac OS X, Pocket PC, Palm OS, Symbian,
etc.
• Supports serializing of application’s native C and C++ data structures, which allows you to
save and load of XML serialized data structures to and from files.
12
• Selective input and output buffering is used to increase efficiency, but full message buffering
to determine HTTP message length is not used. Instead, a three-phase serialization method is
used to determine message length. As a result, large data sets such as base64-encoded images
can be transmitted with or without DIME attachments by small-memory devices such as
PDAs.
• Supports C++ single class inheritance, dynamic binding, overloading, arbitrary pointer structures such as lists, trees, graphs, cyclic graphs, fixed-size arrays, (multi-dimensional) dynamic arrays, enumerations, built-in XSD Schema types including base64Binary encoding,
and hexBinary encoding.
• No need to rewrite existing C/C++ applications for Web service deployment. However, parts
of an application that use unions, pointers to sequences of elements in memory, and void* need
to be modified, but only if the data structures that adopt them are required to be serialized
or deserialized as part of a remote method invocation.
• Three-phase marshalling: 1) analysis of pointers, single-reference, multi-reference, and cyclic
data structures, 2) HTTP message-length determination, and 3) serialization as per SOAP
1.1 encoding style or user-defined encoding styles.
• Two-phase demarshalling: 1) SOAP parsing and decoding, which involves the reconstruction
of multi-reference and cyclic data structures from the payload, and 2) resolution of ”forward”
pointers (i.e. resolution of the forward href attributes in SOAP).
• Full and customizable SOAP Fault processing (client receive and service send).
• Customizable SOAP Header processing (send and receive), which for example enables easy
transaction processing for the service to keep state information.
2
Notational Conventions
The typographical conventions used by this document are:
Sans serif or italics font denotes C and C++ source code, file names, and shell/batch commands.
Bold font denotes C and C++ keywords.
Courier font denotes HTTP header content, HTML, XML, XML Schema content and WSDL
content.
[Optional] denotes an optional construct.
The keywords ”MUST”, ”MUST NOT”, ”REQUIRED”, ”SHALL”, ”SHALL NOT”, ”SHOULD”,
”SHOULD NOT”, ”RECOMMENDED”, ”MAY”, and ”OPTIONAL” in this document are to be
interpreted as described in RFC-2119.
13
3
Differences Between gSOAP Versions 2.4 (and Earlier) and 2.5
To comply with WS-I Basic Profile 1.0a, gSOAP 2.5 and higher adopts SOAP document/literal
by default. There is no need for concern, because the WSDL parser wsdl2h automatically takes
care of the differences when you provide a WSDL document, because SOAP RPC encoding, literal,
and document style are supported. A new soapcpp2 compiler option was added -e for backward
compatibility with gSOAP 2.4 and earlier to adopt SOAP RPC encoding by default in case you want
to develop a service that uses SOAP encoding. You can also use the gSOAP compiler directives to
specify SOAP encoding for individual operarations, when desired.
4
Differences Between gSOAP Versions 2.1 (and Earlier) and 2.2
You should read this section only if you are upgrading from gSOAP 2.1 to 2.2 and later.
Run-time options and flags have been changed to enable separate recv/send settings for transport,
content encodings, and mappings. The flags are divided into four classes: transport (IO), content
encoding (ENC), XML marshalling (XML), and C/C++ data mapping (C). The old-style flags
soap disable X and soap enable X, where X is a particular feature, are deprecated. See Section 8.12
for more details.
5
Differences Between gSOAP Versions 1.X and 2.X
You should read this section only if you are upgrading from gSOAP 1.X to 2.X.
gSOAP versions 2.0 and later have been rewritten based on versions 1.X. gSOAP 2.0 and later is
thread-safe, while 1.X is not. All files in the gSOAP 2.X distribution are renamed to avoid confusion
with gSOAP version 1.X files:
gSOAP 1.X
soapcpp
soapcpp.exe
stdsoap.h
stdsoap.c
stdsoap.cpp
gSOAP 2.X
soapcpp2
soapcpp2.exe
stdsoap2.h
stdsoap2.c
stdsoap2.cpp
Changing the version 1.X application codes to accommodate gSOAP 2.X does not require a significant amount of recoding. The change to gSOAP 2.X affects all functions defined in stdsoap2.c[pp] (the
gSOAP runtime environment API) and the functions in the sources generated by the gSOAP compiler (the gSOAP RPC+marshalling API). Therefore, clients and services developed with gSOAP
1.X need to be modified to accommodate a change in the calling convention used in 2.X: In 2.X, all
gSOAP functions (including the remote method proxy routines) take an additional parameter which
is an instance of the gSOAP runtime environment that includes file descriptors, tables, buffers, and
flags. This additional parameter is always the first parameter of any gSOAP function.
The gSOAP runtime environment is stored in a struct soap type. A struct was chosen to support
application development in C without the need for a separate gSOAP implementation. An objectoriented approach with a class for the gSOAP runtime environment would have prohibited the
14
implementation of pure C applications. Before a client can invoke remote methods or before a
service can accept requests, a runtime environment need to be allocated and initialized. Three new
functions are added to gSOAP 2.X:
Function
soap init(struct soap *soap)
struct soap *soap new()
struct soap *soap copy(struct soap *soap)
Description
Initializes a static or stack-allocated environment (required
only once)
Allocates, initializes, and returns a pointer to a runtime
environment
Allocates a new runtime environment and copies contents
of the environment such that the new environment does
not share any data with the original environment
An environment can be reused as many times as necessary and does not need to be reinitialized in
doing so. A dynamically allocated environment is deallocated with soap free.
A new environment is only required for each new thread to guarantee exclusive access to a new
runtime environment by each thread. For example, the following code stack-allocates the runtime
environment which is used for multiple remote method calls:
int main()
{
struct soap soap;
...
soap init(&soap); // initialize runtime environment
...
soap call ns method1(&soap, ...); // make a remote call
...
soap call ns method2(&soap, ...); // make another remote call
...
soap destroy(&soap); // remove deserialized class instances (C++ only)
soap end(&soap); // clean up and remove deserialized data
soap done(&soap); // detach environment (last use and no longer in scope)
...
}
The runtime environment can also be heap allocated:
int main()
{
struct soap *soap;
...
soap = soap new(); // allocate and initialize runtime environment
if (!soap) // couldn’t allocate: stop
...
soap call ns method1(soap, ...); // make a remote call
...
soap call ns method2(soap, ...); // make another remote call
...
soap destroy(soap); // remove deserialized class instances (C++ only)
15
soap end(soap); // clean up and remove deserialized data
soap free(soap); // detach and free runtime environment
}
A service need to allocate and initialize an environment before calling soap serve:
int main()
{
struct soap soap;
soap init(&soap);
soap serve(&soap);
}
Or alternatively:
int main()
{
soap serve(soap new());
}
The soap serve dispatcher handles one request or multiple requests when HTTP keep-alive is enabled
(with the SOAP IO KEEPALIVE flag see Section 18.11).
A service can use multi-threading to handle requests while running some other code that invokes
remote methods:
int main()
{
struct soap soap1, soap2;
pthread t tid;
...
soap init(&soap1);
if (soap bind(&soap1, host, port, backlog) < 0) exit(1);
if (soap accept(&soap1) < 0) exit(1);
pthread create(&tid, NULL, (void*(*)(void*))soap serve, (void*)&soap1);
...
soap init(&soap2);
soap call ns method(&soap2, ...); // make a remote call
...
soap end(&soap2);
...
pthread join(tid, NULL); // wait for thread to terminate
soap end(&soap1); // release its data
}
In the example above, two runtime environments are required. In comparison, gSOAP 1.X statically
allocates the runtime environment, which prohibits multi-threading (only one thread can invoke
remote methods and/or accept requests due to the single runtime environment).
Section 7.2.4 presents a multi-threaded stand-alone Web Service that handles multiple SOAP requests by spawning a thread for each request.
16
6
Interoperability
gSOAP interoperability has been verified with the following SOAP implementations and toolkits:
Apache 2.2
Apache Axis
ASP.NET
Cape Connect
Delphi
easySOAP++
eSOAP
Frontier
GLUE
Iona XMLBus
kSOAP
MS SOAP
Phalanx
SIM
SOAP::Lite
SOAP4R
Spray
SQLData
Wasp Adv.
Wasp C++
White Mesa
xSOAP
ZSI
4S4C
17
7
Quick User Guide
This user guide offers a quick way to get started with gSOAP. This section requires a basic understanding of the SOAP 1.1 protocol and some familiarity with C and/or C++. In principle, SOAP
clients and SOAP Web services can be developed in C and C++ with the gSOAP compiler without
a detailed understanding of the SOAP protocol when gSOAP client-server applications are built as
an ensamble and only communicate within this group (i.e. meaning that you don’t have to worry
about interoperability with other SOAP implementations). This section is intended to illustrate
the implementation of gSOAP Web services and clients that connect to and interoperate with other
SOAP implementations such as Apache Axis, SOAP::Lite, and .NET. This requires some details of
the SOAP and WSDL protocols to be understood.
7.1
How to Use the gSOAP Stub and Skeleton Compiler to Build SOAP Clients
In general, the implementation of a SOAP client application requires a stub routine for each remote
method that the client application needs to invoke. The primary stub’s responsibility is to marshall
the parameter data, send the request with the parameters to the designated SOAP service over
the wire, to wait for the response, and to demarshall the parameter data of the response when it
arrives. The client application invokes the stub routine for a remote method as if it would invoke
a local method. To write a stub routine in C or C++ by hand is a tedious task, especially if
the input and/or output parameters of a remote method contain elaborate data structures such as
records, arrays, and graphs. Fortunately, the gSOAP ’wsdl2h’ WSDL parser and ’soapcpp2’ stub
and skeleton compiler automate the development of Web service client and server applications.
The gSOAP stub and skeleton compiler is a preprocessor that generates the necessary C++
sources to build SOAP C++ clients. The input to the gSOAP stub and skeleton compiler consists
of a standard C/C++ header file. The header file can be generated from a WSDL (Web Service
Description Language) documentation of a service with the gSOAP WSDL parser.
Consider the following command (entered at the command prompt):
$ wsdl2h -o quote.h http://services.xmethods.net/soap/urn:xmethods-delayed-quotes.wsdl
This generates the file quote.h in C++ format from the WSDL at the specified URL.
To generate a header file to develop a pure C client application, issue the command: Consider the
following command (entered at the command prompt):
$ wsdl2h -c -o quote.h http://services.xmethods.net/soap/urn:xmethods-delayed-quotes.wsdl
For more details on the WSDL parser and its options, see 7.2.10.
The quote.h header file is then processed by the gSOAP compiler to generate the stubs to develop
client applications (and skeletons to develop a service).
The SOAP service methods are specified in the header file as function prototypes. Stub routines
in C/C++ source form are automatically generated by the gSOAP compiler for these function
prototypes of remote methods. The resulting stub routines allow C and C++ client applications
to seamlessly interact with existing SOAP Web services.
18
The gSOAP stub and skeleton compiler also generates skeleton routines for each of the remote
methods specified in the header file. The skeleton routines can be readily used to implement one
or more of the remote methods in a new SOAP Web service. These skeleton routines are not used
for building SOAP clients in C++, although they can be used to build mixed SOAP client/server
applications (peer applications).
The input and output parameters of a SOAP service method may be simple data types or compound
data types, either generated by the WSDL parser or specified by hand. The gSOAP stub and
skeleton compiler automatically generates serializers and deserializers for the data types to
enable the generated stub routines to encode and decode the contents of the parameters of the
remote methods in XML.
7.1.1
Example
The getQuote remote method of XMethods Delayed Stock Quote service (defined in the quote.h file
obtained with the ’wsdl2h’ WSDL parser) provides a delayed stock quote for a given ticker name.
The WSDL description of the XMethods Delayed Stock Quote service provides the following details:
Endpoint URL:
SOAP action:
Remote method namespace:
Remote method name:
Input parameter:
Output parameter:
http://services.xmethods.net:80/soap
”” (2 quotes)
urn:xmethods-delayed-quotes
getQuote
symbol of type xsd:string
Result of type xsd:float
The following getQuote.h header file for C is created from the WSDL description with the WSDL
parser (the actual contents may vary depending on the ’wsdl2h’ release version and the options
used to determine the output):
//gsoap ns1 service name: net DOTxmethods DOTservices DOTstockquote DOTStockQuoteBinding
//gsoap ns1 service type: net DOTxmethods DOTservices DOTstockquote DOTStockQuotePortType
//gsoap ns1 service port: http://66.28.98.121:9090/soap
//gsoap ns1 service namespace: urn:xmethods-delayed-quotes
//gsoap ns1 service documentation: Definitions generated by the gSOAP WSDL parser 1.0
// Service net.xmethods.services.stockquote.StockQuoteService : net.xmethods.services.stockquote.StockQuote
web service
//gsoap
//gsoap
//gsoap
int ns1
ns1 service method-style: getQuote rpc
ns1 service method-encoding: getQuote http://schemas.xmlsoap.org/soap/encoding/
ns1 service method-action: getQuote urn:xmethods-delayed-quotes#getQuote
getQuote(char *symbol, float &Result);
The header file essentially specifies the service details in C/C++ with directives for the gSOAP
compiler. The remote method is declared as a ns1 getQuote function prototype which specifies all
of the necessary details for the gSOAP compiler to generate the stub routine for a client application
to interact with the Delayed Stock Quote service.
The Delayed Stock Quote service description requires that the input parameter of the getQuote
remote method is a symbol parameter of type string. The description also indicates that the Result
19
output parameter is a floating point number that represents the current unit price of the stock
in dollars. The gSOAP compiler uses the convention the last parameter of the function prototype
must be the output parameter of the remote method, which is required to be passed by reference
using the reference operator (&) or by using a pointer type. All other parameters except the last are
input parameters of the remote method, which are required to be passed by value or passed using
a pointer to a value (by reference is not allowed). The function prototype associated with a remote
method is required to return an int, whose value indicates to the caller whether the connection to
a SOAP Web service was successful or resulted in an exception, see Section 9.2 for the error codes.
The use of the namespace prefix ns1 in the remote method name in the function prototype
declaration is discussed in detail in 7.1.2. Basically, a namespace prefix is distinguished by a pair
of underscores in the function name, as in ns1 getQuote where ns1 is the namespace prefix and
getQuote is the remote method name. (A single underscore in an identifier name will be translated
into a dash in XML, because dashes are more frequently used in XML compared to underscores,
see Section 9.3.)
The gSOAP compiler is invoked from the command line with:
soapcpp2 getQuote.h
The compiler generates the stub routine for the getQuote remote method specified in the getQuote.h
header file. This stub routine can be called by a client program at any time to request a stock
quote from the Delayed Stock Quote service. The interface to the generated stub routine is the
following function prototype generated by the gSOAP compiler:
int soap call ns1 getQuote(struct soap *soap, char *URL, char *action, char *symbol, float
&Result);
The stub routine is saved in soapClient.cpp. The file soapC.cpp contains the serializer and deserializer routines for the data types used by the stub. You can use option -c for the soapcpp2 compiler
to generate pure C code.
Note that the parameters of the soap call ns1 getQuote function are identical to the ns1 getQuote
function prototype with three additional input parameters: soap must be a valid pointer to a
gSOAP runtime environment, URL is the SOAP Web service endpoint URL passed as a string,
and action is a string that denotes the SOAP action required by the Web service. Note that the
XMethods Delayed Stock Quote service endpoint URL is http://66.28.98.121:9090/soap and the
SOAP action required is "" (two quotes). You can change the endpoint and action dynamically.
The endpoint and action are the second and third parameters of the soap call ns1 getQuote. When
NULL, the values specified in the header file will be used.
The following example mixed C/C++ client program invokes the stub to retrieve the latest IBM
stock quote from the XMethods Delayed Stock Quote service:
#include "soapH.h" // obtain the generated stub
#include "net_DOT_xmethods_DOT_services_DOT_stockquote_DOT_StockQuoteBinding.nsmap"
// obtain the namespace mapping table
int main()
{
struct soap soap; // gSOAP runtime environment
20
float quote;
soap init(&soap); // initialize runtime environment (only once)
if (soap call ns1 getQuote(&soap, NULL, NULL, "IBM", &quote) == SOAP OK)
std::cout << ”Current IBM Stock Quote = ” << quote << std::endl;
else // an error occurred
soap print fault(&soap, stderr); // display the SOAP fault message on the stderr stream
soap destroy(&soap); // delete deserialized class instances (for C++ only)
soap end(&soap); // remove deserialized data and clean up
soap done(&soap); // detach the gSOAP environment
return 0;
}
When successful, the stub returns SOAP OK and quote contains the latest stock quote. Otherwise,
an error occurred and the SOAP fault is displayed with the soap print fault function.
The gSOAP compiler also generates a proxy class for C++ client applications. This generated
proxy class can be included into a client application together with the generated namespace table
as shown in this example:
#include "soapnet_DOT_xmethods_DOT_services_DOT_stockquote_DOT_StockQuoteBindingProxy.h"
// get proxy
#include "net_DOT_xmethods_DOT_services_DOT_stockquote_DOT_StockQuoteBinding.nsmap"
// obtain the namespace mapping table
int main()
{
net q; // ”net” is the proxy class with a name that is the short name of the service
float r;
if (q.ns1 getQuote(”IBM”, r) == SOAP OK)
std::cout << r << std::endl;
else
soap print fault(q.soap, stderr);
return 0;
}
The proxy class constructor allocates and initializes a gSOAP environment for the instance. All
the HTTP and SOAP/XML processing is hidden and performed on the background. Note that you
can change the name of the service in the header file generated by the WSDL parser. The name is
extracted from the WSDL content and may not always be in a short format. Feel free to change
the entry
//gsoap ns1 service name: net DOT xmethods DOT services DOT stockquote DOT StockQuoteBinding
to use a more suitable name. The name will control the file name of the proxy class file and the
XML namespace mapping table.
The following functions can be used to explicitly setup a gSOAP runtime environment (struct soap):
21
Function
soap init(struct soap *soap)
soap init1(struct soap *soap, soap mode iomode)
soap init2(struct soap *soap, soap mode imode, soap mode omode)
struct soap *soap new()
struct soap *soap new1(soap mode iomode)
struct soap *soap new2(soap mode imode, soap mode omode)
struct soap *soap copy(struct soap *soap)
soap done(struct soap *soap)
soap free(struct soap *soap)
Description
Initializes a static/stack-allocated runtime env
Initializes a runtime environment and set in/ou
Initializes a runtime environment and set sep
mode flags
Allocates, initializes, and returns a pointer t
environment
Allocates, initializes, and returns a pointer t
environment and set in/out mode flags
Allocates, initializes, and returns a pointer t
environment and set separate in/out mode fla
Allocates a new runtime environment and cop
of the source environment such that the new e
does not share data with the source environm
Reset, close communications, and remove call
Reset and deallocate the environment cr
soap new or soap copy
An environment can be reused as many times as necessary for client-side remote calls and does
not need to be reinitialized in doing so. A new environment is required for each new thread to
guarantee exclusive access to runtime environments by threads. Also the use of any client calls
within an active service method requires a new environment.
When the example client application is invoked, the SOAP request is performed by the stub routine
soap call ns1 getQuote, which generates the following SOAP RPC request message:
POST /soap HTTP/1.1
Host: services.xmethods.net
Content-Type: text/xml
Content-Length: 529
SOAPAction: ""
<?xml version="1.0" encoding="UTF-8"?>
<SOAP-ENV:Envelope xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:SOAP-ENC="http://schemas.xmlsoap.org/soap/encoding/"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:xsd="http://www.w3.org/2001/XMLSchema"
xmlns:ns1="urn:xmethods-delayed-quotes"
SOAP-ENV:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/">
<SOAP-ENV:Body>
<ns1:getQuote>
<symbol>IBM</symbol>
</ns1:getQuote>
</SOAP-ENV:Body>
</SOAP-ENV:Envelope>
The XMethods Delayed Stock Quote service responds with the SOAP response message:
HTTP/1.1 200 OK
Date: Sat, 25 Aug 2001 19:28:59 GMT
Content-Type: text/xml
22
Server: Electric/1.0
Connection: Keep-Alive
Content-Length: 491
<?xml version="1.0" encoding="UTF-8"?>
<soap:Envelope xmlns:soap="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:xsd="http://www.w3.org/2001/XMLSchema"
xmlns:soapenc="http://schemas.xmlsoap.org/soap/encoding/"
soap:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/">
<soap:Body>
<n:getQuoteResponse xmlns:n="urn:xmethods-delayed-quotes">
<Result xsi:type="xsd:float">41.81</Result>
</n:getQuoteResponse>
</soap:Body>
</soap:Envelope>
The server’s SOAP RPC response is parsed by the stub. The stub routine further demarshalls the
data of Result element of the SOAP response and stores it in the quote parameter of soap call ns1 getQuote.
A client program can invoke a remote method at any time and multiple times if necessary. Consider
for example:
...
struct soap soap;
float quotes[3]; char *myportfolio[] = {"IBM", "MSDN"};
soap init(&soap); // need to initialize only once
for (int i = 0; i < 3; i++)
if (soap call ns1 getQuote(&soap, "http://services.xmethods.net:80/soap", "", myportfolio[i], &quotes[i]) != SOAP OK)
break;
if (soap.error) // an error occurred
soap print fault(&soap, stderr);
soap end(&soap); // clean up all deserialized data
...
This client composes an array of stock quotes by calling the ns1 getQuote stub routine for each
symbol in a portfolio array.
This example demonstrated how easy it is to build a SOAP client with gSOAP once the details of
a Web service are available in the form of a WSDL document.
7.1.2
Namespace Considerations
The declaration of the ns1 getQuote function prototype (discussed in the previous section) uses
the namespace prefix ns1 of the remote method namespace, which is distinguished by a pair
of underscores in the function name to separate the namespace prefix from the remote method
name. The purpose of a namespace prefix is to associate a remote method name with a service
in order to prevent naming conflicts, e.g. to distinguish identical remote method names used by
different services.
23
Note that the XML response of the XMethods Delayed Stock Quote service example uses the
namespace prefix n which is bound to the namespace name urn:xmethods-delayed-quotes
through the xmlns:n="urn:xmethods-delayed-quotes binding. The use of namespace prefixes and
namespace names is also required to enable SOAP applications to validate the content of SOAP
messages. The namespace name in the service response is verified by the stub routine by using the
information supplied in a namespace mapping table that is required to be part of gSOAP client
and service application codes. The table is accessed at run time to resolve namespace bindings, both
by the generated stub’s data structure serializer for encoding the client request and by the generated
stub’s data structure deserializer to decode and validate the service response. The namespace
mapping table should not be part of the header file input to the gSOAP stub and skeleton compiler.
Service details including namespace bindings may be provided with gSOAP directives in a header
file, see Section 18.2.
The namespace mapping table for the Delayed Stock Quote client is:
struct Namespace namespaces[] =
{ // {”ns-prefix”, ”ns-name”}
{”SOAP-ENV”, ”http://schemas.xmlsoap.org/soap/envelope/”}, // MUST be first
{”SOAP-ENC”, ”http://schemas.xmlsoap.org/soap/encoding/”}, // MUST be second
{”xsi”, ”http://www.w3.org/2001/XMLSchema-instance”}, // MUST be third
{”xsd”, ”http://www.w3.org/2001/XMLSchema”}, // 2001 XML Schema
{”ns1”, ”urn:xmethods-delayed-quotes”}, // given by the service description
{NULL, NULL} // end of table
};
The first four namespace entries in the table consist of the standard namespaces used by the SOAP
1.1 protocol. In fact, the namespace mapping table is explicitly declared to enable a programmer
to specify the SOAP encoding style and to allow the inclusion of namespace-prefix with namespacename bindings to comply to the namespace requirements of a specific SOAP service. For example,
the namespace prefix ns1, which is bound to urn:xmethods-delayed-quotes by the namespace mapping table shown above, is used by the generated stub routine to encode the getQuote request. This
is performed automatically by the gSOAP compiler by using the ns1 prefix of the ns1 getQuote
method name specified in the getQuote.h header file. In general, if a function name of a remote
method, struct name, class name, enum name, or field name of a struct or class has a pair of
underscores, the name has a namespace prefix that must be defined in the namespace mapping
table.
The namespace mapping table will be output as part of the SOAP Envelope by the stub routine.
For example:
...
<SOAP-ENV:Envelope xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:SOAP-ENC="http://schemas.xmlsoap.org/soap/encoding/"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:xsd="http://www.w3.org/2001/XMLSchema"
xmlns:ns1="urn:xmethods-delayed-quotes"
SOAP-ENV:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/">
...
The namespace bindings will be used by a SOAP service to validate the SOAP request.
24
7.1.3
Example
The incorporation of namespace prefixes into C++ identifier names is necessary to distinguish
remote methods that share the same name but are provided by separate Web services and/or
organizations. Consider for example:
// Contents of file ”getQuote.h”:
int ns1 getQuote(char *symbol, float &Result);
int ns2 getQuote(char *ticker, char *&quote);
Recall that the namespace prefix is always separated from the name of a remote method by a pair
of underscores ( ).
This example enables a client program to connect to a (hypothetical) Stock Quote service with
remote methods that can only be distinguished by their namespaces. Consequently, two different
namespace prefixes had to be used as part of the remote method names.
The namespace prefix convention can also be applied to class declarations that contain SOAP
compound values that share the same name but have different namespaces that refer to different
XML Schemas. For example:
class e Address // an electronic address
{
char *email;
char *url;
};
class s Address // a street address
{
char *street;
int number;
char *city;
};
The namespace prefix is separated from the name of a data type by a pair of underscores ( ).
An instance of e Address is encoded by the generated serializer for this type as an Address element
with namespace prefix e:
<e:Address xsi:type="e:Address">
<email xsi:type="string">me@home</email>
<url xsi:type="string">www.me.com</url>
</e:Address>
While an instance of s Address is encoded by the generated serializer for this type as an Address
element with namespace prefix s:
<s:Address xsi:type="s:Address">
<street xsi:type="string">Technology Drive</street>
<number xsi:type="int">5</number>
<city xsi:type="string">Softcity</city>
</s:Address>
25
The namespace mapping table of the client program must have entries for e and s that refer to the
XML Schemas of the data types:
struct Namespace namespaces[] =
{ ...
{”e”, ”http://www.me.com/schemas/electronic-address”},
{”s”, ”http://www.me.com/schemas/street-address”},
...
This table is required to be part of the client application to allow access by the serializers and
deserializers of the data types at run time.
7.1.4
How to Generate C++ Client Proxy Classes
Proxy classes for C++ client applications are automatically generated by the gSOAP compiler.
There is a new and improved code generation capability for proxy classes, which is activated with
the soapcpp2 -i option. These new proxy classes are derived from the soap structure, have a cleaner
interface and offer more capabilites.
To illustrate the generation of a “standard” (old) proxy class, the getQuote.h header file example of
the previous section is augmented with the appropriate directives to enable the gSOAP compiler to
generate the proxy class. Similar directives are included in the header file by the WSDL importer.
// Content of file "getQuote.h":
//gsoap ns1 service name: Quote
//gsoap ns1 service location: http://services.xmethods.net/soap
//gsoap ns1 service namespace: urn:xmethods-delayed-quotes
//gsoap ns1 service style: rpc
//gsoap ns1 service encoding: encoded
//gsoap ns1 service method-action: getQuote ””
int ns1 getQuote(char *symbol, float &Result);
The first three directives provide the service name which is used to name the proxy class, the
service location (endpoint), and the schema. The forth and fifth directives specify that SOAP
RPC encoding is used, which is required by this service. The last directive defines the optional
SOAPAction, which is a string associated with SOAP 1.1 operations. This directive must be
provided for each remote method when the SOAPAction is required. Compilation of this header
file with the gSOAP compiler soapcpp2 creates a new file soapQuoteProxy.h with the following contents:
#include ”soapH.h”
class Quote
{ public:
struct soap *soap;
const char *endpoint;
Quote() { soap = soap new(); endpoint = ”http://services.xmethods.net/soap”; };
˜Quote() { if (soap) { soap destroy(soap); soap end(soap); soap free(soap); }};
int getQuote(char *symbol, float &Result) { return soap ? soap call ns1 getQuote(soap, endpoint, ””, symbol, Result) : SOAP EOM; };
};
26
The gSOAP environment and endpoint are declared public to enable access for run-time customization.
This generated proxy class can be included into a client application together with the generated
namespace table as shown in this example:
#include ”soapQuoteProxy.h” // get proxy
#include ”Quote.nsmap” // get namespace bindings
int main()
{
Quote q;
float r;
if (q.ns1 getQuote(”IBM”, r) == SOAP OK)
std::cout << r << std::endl;
else
soap print fault(q.soap, stderr);
return 0;
}
The Quote constructor allocates and initializes a gSOAP environment for the instance. All the
HTTP and SOAP/XML processing is hidden and performed on the background.
You can use soapcpp2 compiler option -n together with -p to create a local namespaces table to
avoid link conflicts when you need multiple namespace tables or need to combine multiple clients,
see also Sections 8.1 and 18.34, and you can use a C++ code namespace to create a namespace
qualified proxy class, see Section 18.33.
Don’t forget to try the soapcpp2 -i option to generate proxy classes derived from the base soap
structure. In addition, these classes offer more functionality.
7.1.5
XSD Type Encoding Considerations
Many SOAP services require the explicit use of XML Schema types in the SOAP payload. The
default encoding, which is also adopted by the gSOAP compiler, assumes SOAP RPC encoding
which only requires the use of types to handle polymorphic cases. Nevertheless, the use of XSD
typed messages is advised to improve interoperability. XSD types are introduced with typedef
definitions in the header file input to the gSOAP compiler. The type name defined by a typedef
definition corresponds to an XML Schema type (XSD type). For example, the following typedef
declarations define various built-in XSD types implemented as primitive C/C++ types:
// Contents of header file:
...
typedef char *xsd string; // encode xsd string value as the xsd:string schema type
typedef char *xsd anyURI; // encode xsd anyURI value as the xsd:anyURI schema type
typedef float xsd float; // encode xsd float value as the xsd:float schema type
typedef long xsd int; // encode xsd int value as the xsd:int schema type
typedef bool xsd boolean; // encode xsd boolean value as the xsd:boolean schema type
typedef unsigned long long xsd positiveInteger; // encode xsd positiveInteger value as the
xsd:positiveInteger schema type
...
27
This simple mechanism informs the gSOAP compiler to generate serializers and deserializers that
explicitly encode and decode the primitive C++ types as built-in primitive XSD types when the
typedefed type is used in the parameter signature of a remote method (or when used nested within
structs, classes, and arrays). At the same time, the use of typedef does not force any recoding of a
C++ client or Web service application as the internal C++ types used by the application are not
required to be changed (but still have to be primitive C++ types, see Section 10.3.2 for alternative
class implementations of primitive XSD types which allows for the marshalling of polymorphic
primitive types).
7.1.6
Example
Reconsider the getQuote example, now rewritten with explicit XSD types to illustrate the effect:
// Contents of file ”getQuote.h”:
typedef char *xsd string;
typedef float xsd float;
int ns1 getQuote(xsd string symbol, xsd float &Result);
This header file is compiled by the gSOAP stub and skeleton compiler and the compiler generates
source code for the function soap call ns1 getQuote, which is identical to the “old” proxy:
int soap call ns1 getQuote(struct soap *soap, char *URL, char *action, char *symbol, float
&Result);
The client application does not need to be rewritten and can still call the proxy using the “old”
parameter signature. In contrast to the previous implementation of the stub however, the encoding
and decoding of the data types by the stub has been changed to explicitly use the XSD types in
the message payload.
For example, when the client application calls the proxy, the proxy produces a SOAP request with
an xsd:string:
...
<SOAP-ENV:Body>
<ns1:getQuote><symbol xsi:type="xsd:string">IBM</symbol>
</ns1:getQuote>
</SOAP-ENV:Body>
...
The service response is:
...
<soap:Body>
<n:getQuoteResponse xmlns:n="urn:xmethods-delayed-quotes">
<Result xsi:type="xsd:float">41.81</Result>
</n:getQuoteResponse>
</soap:Body>
...
28
The validation of this service response by the stub routine takes place by matching the namespace names (URIs) that are bound to the xsd namespace prefix. The stub also expects the
getQuoteResponse element to be associated with URI urn:xmethods-delayed-quotes through the
binding of the namespace prefix ns1 in the namespace mapping table. The service response uses
namespace prefix n for the getQuoteResponse element. This namespace prefix is bound to the same
URI urn:xmethods-delayed-quotes and therefore the service response is assumed to be valid. The
response is rejected and a SOAP fault is generated when the URIs do not match.
7.1.7
How to Change the Response Element Name
There is no standardized convention for the response element name in a SOAP response message,
although it is recommended that the response element name is the method name ending with
“Response”. For example, the response element of getQuote is getQuoteResponse.
The response element name can be specified explicitly using a struct or class declaration in the
header file. The struct or class name represents the SOAP response element name used by the
service. Consequently, the output parameter of the remote method must be declared as a field of
the struct or class. The use of a struct or a class for the service response is fully SOAP 1.1 compliant.
In fact, the absence of a struct or class indicates to the gSOAP compiler to automatically generate
a struct for the response which is internally used by a stub.
7.1.8
Example
Reconsider the getQuote remote method specification which can be rewritten with an explicit declaration of a SOAP response element as follows:
// Contents of ”getQuote.h”:
typedef char *xsd string;
typedef float xsd float;
struct ns1 getQuoteResponse {xsd float Result;};
int ns1 getQuote(xsd string symbol, struct ns1 getQuoteResponse &r);
The SOAP request is the same as before:
...
<SOAP-ENV:Body>
<ns1:getQuote><symbol xsi:type="xsd:string">IBM</symbol>
</ns1:getQuote>
</SOAP-ENV:Body>
...
The difference is that the service response is required to match the specified getQuoteResponse name
and its namespace URI:
...
<soap:Body>
<n:getQuoteResponse xmlns:n=’urn:xmethods-delayed-quotes’>
29
<Result xsi:type=’xsd:float’>41.81</Result>
</n:getQuoteResponse>
</soap:Body>
...
This use of a struct or class enables the adaptation of the default SOAP response element name
and/or namespace URI when required.
Note that the struct (or class) declaration may appear within the function prototype declaration.
For example:
// Contents of ”getQuote.h”:
typedef char *xsd string;
typedef float xsd float;
int ns1 getQuote(xsd string symbol, struct ns1 getQuoteResponse {xsd float Result;} &r);
This example combines the declaration of the response element of the remote method with the
function prototype of the remote method.
7.1.9
How to Specify Multiple Output Parameters
The gSOAP stub and skeleton compiler uses the convention that the last parameter of the
function prototype declaration of a remove method in a header file is also the only single output
parameter of the method. All other parameters are considered input parameters of the remote
method. To specify a remote method with multiple output parameters, a struct or class must
be declared for the remote method response, see also 7.1.7. The fields of the struct or class are the
output parameters of the remote method. Both the order of the input parameters in the function
prototype and the order of the output parameters (the fields in the struct or class) is not significant.
However, the SOAP 1.1 specification states that input and output parameters may be treated as
having anonymous parameter names which requires a particular ordering, see Section 7.1.13.
7.1.10
Example
As an example, consider a hypothetical remote method getNames with a single input parameter SSN
and two output parameters first and last. This can be specified as:
// Contents of file ”getNames.h”:
int ns3 getNames(char *SSN, struct ns3 getNamesResponse {char *first; char *last;} &r);
The gSOAP stub and skeleton compiler takes this header file as input and generates source code
for the function soap call ns3 getNames. When invoked by a client application, the proxy produces
the SOAP request:
...
<SOAP-ENV:Envelope ...
...
<ns3:getNames>
xmlns:ns3="urn:names" ...>
30
<SSN>999 99 9999</SSN>
</ns3:getNames>
...
The response by a SOAP service could be:
...
<m:getNamesResponse xmlns:m="urn:names">
<first>John</first>
<last>Doe</last>
</m:getNamesResponse>
...
where first and last are the output parameters of the getNames remote method of the service.
As another example, consider a remote method copy with an input parameter and an output parameter with identical parameter names (this is not prohibited by the SOAP 1.1 protocol). This
can be specified as well using a response struct:
// Content of file ”copy.h”:
int X rox copy name(char *name, struct X rox copy nameResponse {char *name;} &r);
The use of a struct or class for the remote method response enables the declaration of remote
methods that have parameters that are passed both as input and output parameters.
The gSOAP compiler takes the copy.h header file as input and generates the soap call X rox copy name
proxy. When invoked by a client application, the proxy produces the SOAP request:
...
<SOAP-ENV:Envelope ...
...
<X-rox:copy-name>
<name>SOAP</name>
</X-rox:copy-name>
...
xmlns:X-rox="urn:copy" ...>
The response by a SOAP copy service could be something like:
...
<m:copy-nameResponse xmlns:m="urn:copy">
<name>SOAP</name>
</m:copy-nameResponse>
...
The name will be parsed and decoded by the proxy and returned in the name field of the struct
X rox copy nameResponse &r parameter.
7.1.11
How to Specify Output Parameters With struct/class Compound Data Types
If the single output parameter of a remote method is a complex data type such as a struct or class
it is necessary to specify the response element of the remote method as a struct or class at all
31
times. Otherwise, the output parameter will be considered the response element (!), because of
the response element specification convention used by gSOAP, as discussed in 7.1.7.
7.1.12
Example
This is is best illustrated with an example. The Flighttracker service by ObjectSpace provides
real time flight information for flights in the air. It requires an airline code and flight number as
parameters. The remote method name is getFlightInfo and the method has two string parameters:
the airline code and flight number, both of which must be encoded as xsd:string types. The method
returns a getFlightResponse response element with a return output parameter that is of complex type
FlightInfo. The type FlightInfo is represented by a class in the header file, whose field names correspond
to the FlightInfo accessors:
// Contents of file ”flight.h”:
typedef char *xsd string;
class ns2 FlightInfo
{
public:
xsd string airline;
xsd string flightNumber;
xsd string altitude;
xsd string currentLocation;
xsd string equipment;
xsd string speed;
};
struct ns1 getFlightInfoResponse {ns2 FlightInfo return;};
int ns1 getFlightInfo(xsd string param1, xsd string param2, struct ns1 getFlightInfoResponse
&r);
The response element ns1 getFlightInfoResponse is explicitly declared and it has one field: return of
type ns2 FlightInfo. Note that return has a trailing underscore to avoid a name clash with the return
keyword, see Section 9.3 for details on the translation of C++ identifiers to XML element names.
The gSOAP compiler generates the soap call ns1 getFlightInfo proxy. Here is an example fragment
of a client application that uses this proxy to request flight information:
struct soap soap;
...
soap init(&soap);
...
soap call ns1 getFlightInfo(&soap, "testvger.objectspace.com/soap/servlet/rpcrouter",
"urn:galdemo:flighttracker", "UAL", "184", r);
...
struct Namespace namespaces[] =
{
{”SOAP-ENV”, ”http://schemas.xmlsoap.org/soap/envelope/”},
{”SOAP-ENC”,”http://schemas.xmlsoap.org/soap/encoding/”},
{”xsi”, ”http://www.w3.org/2001/XMLSchema-instance”},
{”xsd”, ”http://www.w3.org/2001/XMLSchema”},
32
{”ns1”, ”urn:galdemo:flighttracker”},
{”ns2”, ”http://galdemo.flighttracker.com”},
{NULL, NULL}
};
When invoked by a client application, the proxy produces the SOAP request:
POST /soap/servlet/rpcrouter HTTP/1.1
Host: testvger.objectspace.com
Content-Type: text/xml
Content-Length: 634
SOAPAction: "urn:galdemo:flighttracker"
<?xml version="1.0" encoding="UTF-8"?>
<SOAP-ENV:Envelope xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:SOAP-ENC="http://schemas.xmlsoap.org/soap/encoding/"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:xsd="http://www.w3.org/2001/XMLSchema"
xmlns:ns1="urn:galdemo:flighttracker"
xmlns:ns2="http://galdemo.flighttracker.com"
SOAP-ENV:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/">
<SOAP-ENV:Body>
<ns1:getFlightInfo xsi:type="ns1:getFlightInfo">
<param1 xsi:type="xsd:string">UAL</param1>
<param2 xsi:type="xsd:string">184</param2>
</ns1:getFlightInfo>
</SOAP-ENV:Body>
</SOAP-ENV:Envelope>
The Flighttracker service responds with:
HTTP/1.1 200 ok
Date: Thu, 30 Aug 2001 00:34:17 GMT
Server: IBM HTTP Server/1.3.12.3 Apache/1.3.12 (Win32)
Set-Cookie: sesessionid=2GFVTOGC30D0LGRGU2L4HFA;Path=/
Cache-Control: no-cache="set-cookie,set-cookie2"
Expires: Thu, 01 Dec 1994 16:00:00 GMT
Content-Length: 861
Content-Type: text/xml; charset=utf-8
Content-Language: en
<?xml version="1.0" encoding="UTF-8"?>
<SOAP-ENV:Envelope xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:xsd="http://www.w3.org/2001/XMLSchema">
<SOAP-ENV:Body>
<ns1:getFlightInfoResponse xmlns:ns1="urn:galdemo:flighttracker"
SOAP-ENV:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/">
<return xmlns:ns2="http://galdemo.flighttracker.com" xsi:type="ns2:FlightInfo">
<equipment xsi:type="xsd:string">A320</equipment>
<airline xsi:type="xsd:string">UAL</airline>
33
<currentLocation xsi:type="xsd:string">188 mi W of Lincoln, NE</currentLocation>
<altitude xsi:type="xsd:string">37000</altitude>
<speed xsi:type="xsd:string">497</speed>
<flightNumber xsi:type="xsd:string">184</flightNumber>
</return>
</ns1:getFlightInfoResponse>
</SOAP-ENV:Body>
</SOAP-ENV:Envelope>
The proxy returns the service response in variable r of type struct ns1 getFlightInfoResponse and this
information can be displayed by the client application with the following code fragment:
cout << r.return .equipment << ” flight ” << r.return .airline << r.return .flightNumber
<< ” traveling ” << r.return .speed << ” mph ” << ” at ” << r.return .altitude
<< ” ft, is located ” << r.return .currentLocation << endl;
This code displays the service response as:
A320 flight UAL184 traveling 497 mph at 37000 ft, is located 188 mi W of Lincoln,
NE
Note: the flight tracker service is no longer available since 9/11/2001. It is kept in the documentation as an example to illustrate the use of structs/classes and response types.
7.1.13
How to Specify Anonymous Parameter Names
The SOAP 1.1 protocol allows parameter names to be anonymous. That is, the name(s) of the
output parameters of a remote method are not strictly required to match a client’s view of the
parameters names. Also, the input parameter names of a remote method are not strictly required to
match a service’s view of the parameter names. Although this convention is likely to be deprecated
in SOAP 1.2, the gSOAP compiler can generate stub and skeleton routines that support anonymous
parameters. Parameter names are implicitly anonymous by omitting the parameter names in the
function prototype of the remote method. For example:
// Contents of ”getQuote.h”:
typedef char *xsd string;
typedef float xsd float;
int ns1 getQuote(xsd string, xsd float&);
To make parameter names explicitly anonymous on the receiving side (client or service), the parameter names should start with an underscore ( ) in the function prototype in the header file.
For example:
// Contents of ”getQuote.h”:
typedef char *xsd string;
typedef float xsd float;
int ns1 getQuote(xsd string symbol, xsd float & return);
34
Or, alternatively with a response struct:
// Contents of ”getQuote.h”:
typedef char *xsd string;
typedef float xsd float;
struct ns1 getQuoteResponse {xsd float return;};
int ns1 getQuote(xsd string symbol, struct ns1 getQuoteResponse &r);
In this example, return is an anonymous output parameter. As a consequence, the service response
to a request made by a client created with gSOAP using this header file specification may include
any name for the output parameter in the SOAP payload. The input parameters may also be
anonymous. This affects the implementation of Web services in gSOAP and the matching of
parameter names by the service.
Caution: when anonymous parameter names are used, the order of the parameters in the function
prototype of a remote method is significant.
7.1.14
How to Specify a Method with No Input Parameters
To specify a remote method that has no input parameters, just provide a function prototype with
one parameter which is the output parameter. However, some C/C++ compilers (notably Visual
C++TM ) will not compile and complain about an empty struct. This struct is generated by gSOAP
to contain the SOAP request message. To fix this, provide one input parameter of type void*
(gSOAP can not serialize void* data). For example:
struct ns3 SOAPService
{
public:
int ID;
char *name;
char *owner;
char *description;
char *homepageURL;
char *endpoint;
char *SOAPAction;
char *methodNamespaceURI;
char *serviceStatus;
char *methodName;
char *dateCreated;
char *downloadURL;
char *wsdlURL;
char *instructions;
char *contactEmail;
char *serverImplementation;
};
struct ArrayOfSOAPService {struct ns3 SOAPService * ptr; int size;};
int ns getAllSOAPServices(void * , struct ArrayOfSOAPService & return);
The ns getAllSOAPServices method has one void* input parameter which is ignored by the serializer
to produce the request message.
35
Most C/C++ compilers allow empty structs and therefore the void* parameter is not required.
7.1.15
How to Specify a Method with No Output Parameters
To specify a remote method that has no output parameters, just provide a function prototype with
a response struct that is empty. For example:
enum ns event { off, on, stand by };
int ns signal(enum ns event in, struct ns signalResponse { } *out);
Since the response struct is empty, no output parameters are specified. Most C/C++ compilers
allow empty structs. For those that don’t, use a void* parameter in the struct. This parameter is
not (de)serialized.
Some SOAP resources refer to SOAP RPC with empty responses as one way SOAP messaging.
However, we refer to one-way massaging by asynchronous explicit send and receive operations as
described in Section 7.3. The latter view of one-way SOAP messaging is also in line with Basic
Profile 1.0.
7.2
How to Use the gSOAP Stub and Skeleton Compiler to Build SOAP Web
Services
The gSOAP stub and skeleton compiler generates skeleton routines in C++ source form for each
of the remote methods specified as function prototypes in the header file processed by the gSOAP
compiler. The skeleton routines can be readily used to implement the remote methods in a new
SOAP Web service. The compound data types used by the input and output parameters of SOAP
remote methods must be declared in the header file, such as structs, classes, arrays, and pointerbased data structures (graphs) that are used as the data types of the parameters of a remote method.
The gSOAP compiler automatically generates serializers and deserializers for the data types to
enable the generated skeleton routines to encode and decode the contents of the parameters of the
remote methods. The gSOAP compiler also generates a remote method request dispatcher routine
that will serve requests by calling the appropriate skeleton when the SOAP service application is
installed as a CGI application on a Web server.
7.2.1
Example
The following example specifies three remote methods to be implemented by a new SOAP Web
service:
// Contents of file ”calc.h”:
typedef double xsd double;
int ns add(xsd double a, xsd double b, xsd double &result);
int ns sub(xsd double a, xsd double b, xsd double &result);
int ns sqrt(xsd double a, xsd double &result);
36
The add and sub methods are intended to add and subtract two double floating point numbers
stored in input parameters a and b and should return the result of the operation in the result output
parameter. The qsrt method is intended to take the square root of input parameter a and to return
the result in the output parameter result. The xsd double type is recognized by the gSOAP compiler
as the xsd:double XSD Schema data type. The use of typedef is a convenient way to associate
primitive C types with primitive XML Schema data types.
To generate the skeleton routines, the gSOAP compiler is invoked from the command line with:
soapcpp2 calc.h
The compiler generates the skeleton routines for the add, sub, and sqrt remote methods specified
in the calc.h header file. The skeleton routines are respectively, soap serve ns add, soap serve ns sub,
and soap serve ns sqrt and saved in the file soapServer.cpp. The generated file soapC.cpp contains
serializers and deserializers for the skeleton. The compiler also generates a service dispatcher:
the soap serve function handles client requests on the standard input stream and dispatches the
remote method requests to the appropriate skeletons to serve the requests. The skeleton in turn
calls the remote method implementation function. The function prototype of the remote method
implementation function is specified in the header file that is input to the gSOAP compiler.
Here is an example Calculator service application that uses the generated soap serve routine to
handle client requests:
// Contents of file ”calc.cpp”:
#include ”soapH.h”
#include <math.h> // for sqrt()
main()
{
soap serve(soap new()); // use the remote method request dispatcher
}
// Implementation of the ”add” remote method:
int ns add(struct soap *soap, double a, double b, double &result)
{
result = a + b;
return SOAP OK;
}
// Implementation of the ”sub” remote method:
int ns sub(struct soap *soap, double a, double b, double &result)
{
result = a - b;
return SOAP OK;
}
// Implementation of the ”sqrt” remote method:
int ns sqrt(struct soap *soap, double a, double &result)
{
if (a >= 0)
{
result = sqrt(a);
return SOAP OK;
}
else
37
return soap receiver fault(soap, ”Square root of negative number”, ”I can only take the square
root of a non-negative number”);
}
// As always, a namespace mapping table is needed:
struct Namespace namespaces[] =
{ // {”ns-prefix”, ”ns-name”}
{”SOAP-ENV”, ”http://schemas.xmlsoap.org/soap/envelope/”},
{”SOAP-ENC”, ”http://schemas.xmlsoap.org/soap/encoding/”},
{”xsi”, ”http://www.w3.org/2001/XMLSchema-instance”},
{”xsd”, ”http://www.w3.org/2001/XMLSchema”},
{”ns”, ”urn:simple-calc”}, // bind ”ns” namespace prefix
{NULL, NULL}
};
Note that the remote methods have an extra input parameter which is a pointer to the gSOAP
runtime environment. The implementation of the remote methods MUST return a SOAP error code.
The code SOAP OK denotes success, while SOAP FAULT denotes an exception with details that can be
defined by the user. The exception description can be assigned to the soap->fault->faultstring string
and details can be assigned to the soap->fault->detail string. This is SOAP 1.1 specific. SOAP
1.2 requires the soap->fault->SOAP ENV Reason and the soap->fault->SOAP ENV Detail strings to
be assigned. Better is to use the soap receiver fault function that allocates a fault struct and sets
the SOAP Fault string and details regardless of the SOAP 1.1 or SOAP 1.2 version used. The
soap receiver fault function returns SOAP FAULT, i.e. an application-specific fault. The fault exception
will be passed on to the client of this service.
This service application can be readily installed as a CGI application. The service description
would be:
Endpoint URL:
SOAP action:
Remote method namespace:
Remote method name:
Input parameters:
Output parameter:
Remote method name:
Input parameters:
Output parameter:
Remote method name:
Input parameter:
Output parameter:
the URL of the CGI application
”” (2 quotes)
urn:simple-calc
add
a of type xsd:double and b of type xsd:double
result of type xsd:double
sub
a of type xsd:double and b of type xsd:double
result of type xsd:double
sqrt
a of type xsd:double
result of type xsd:double or a SOAP Fault
The soapcpp2 compile generates a WSDL file for this service, see Section 7.2.8.
Unless the CGI application inspects and checks the environment variable SOAPAction which contains
the SOAP action request by a client, the SOAP action is ignored by the CGI application. SOAP
actions are specific to the SOAP protocol and provide a means for routing requests and for security
reasons (e.g. firewall software can inspect SOAP action headers to grant or deny the SOAP request.
Note that this requires the SOAP service to check the SOAP action header as well to match it with
the remote method.)
38
The header file input to the gSOAP compiler does not need to be modified to generate client stubs
for accessing this service. Client applications can be developed by using the same header file as
for which the service application was developed. For example, the soap call ns add stub routine is
available from the soapClient.cpp file after invoking the gSOAP compiler on the calc.h header file. As
a result, client and service applications can be developed without the need to know the details of
the SOAP encoding used.
7.2.2
MSVC++ Builds
• Win32 builds need winsock.dll (MS Visual C++ ”wsock32.lib”) To do this in Visual C++
6.0, go to ”Project”, ”settings”, select the ”Link” tab (the project file needs to be selected in
the file view) and add ”wsock32.lib” to the ”Object/library modules” entry.
• Use files with extension .cpp only (don’t mix .c with .cpp).
• Turn pre-compiled headers off.
• When creating a new project, you can specify a custom build step to automatically invoke
the gSOAP compiler on a gSOAP header file. In this way you can incrementally build
a new service by adding new operations and data types to the header file. To specify a
custom build step, select the ”Project” menu item ”Settings” and select the header file in
the File view pane. Select the ”Custom Build” tab and enter ’soapcpp2.exe ”$(inputPath)”’ in
the ”Command” pane. Enter ’soapStub.h soapH.h soapC.cpp soapClient.cpp soapServer.cpp’. Don’t
forget to add the soapXYZProxy.h soapXYZObject.h files that are generated for C++ class proxies
and server objects named XYZ. Click ”OK”. Run the soapcpp2 compiler once to generate
these files (you can simply do this by selecting your header file and select ”Compile”). Add
the files to your project. Each time you make a change to the header file, the project sources
are updated automatically.
• You may want to use the WinInet interface available in the mod gsoap directory of the gSOAP
package to simplify Internet access and deal with encryption, proxies, and authentication.
API instructions are included in the source.
• For the PocketPC, run the wsdl2h WSDL parser with option -s to prevent the generation of
STL code. In addition, time t serialization is not supported, which means that you should add
the following line to typemap.dat indicating a mapping of xsd dateTime to char*: xsd dateTime
= | char* | char*.
7.2.3
How to Create a Stand-Alone gSOAP Service
The deployment of a Web service as a CGI application is an easy means to provide your service
on the Internet. gSOAP services can also run as stand-alone services on any port by utilizing
the built-in HTTP and TCP/IP stacks. The stand-alone services can be run on port 80 thereby
providing Web server capabilities restricted to SOAP RPC.
To create a stand-alone service, only the main routine of the service needs to be modified as follows.
Instead of just calling the soap serve routine, the main routine is changed into:
39
int main()
{
struct soap soap;
int m, s; // master and slave sockets
soap init(&soap);
m = soap bind(&soap, "machine.cs.fsu.edu", 18083, 100);
if (m < 0)
soap print fault(&soap, stderr);
else
{
fprintf(stderr, "Socket connection successful: master socket = %d\n", m);
for (int i = 1; ; i++)
{
s = soap accept(&soap);
if (s < 0)
{
soap print fault(&soap, stderr);
break;
}
fprintf(stderr, "%d: accepted connection from IP=%d.%d.%d.%d socket=%d", i,
(soap.ip>>24)&0xFF, (soap.ip>>16)&0xFF, (soap.ip>>8)&0xFF, soap.ip&0xFF, s);
if (soap serve(&soap) != SOAP OK) // process RPC request
soap print fault(&soap, stderr); // print error
fprintf(stderr, "request served\n");
soap destroy(&soap); // clean up class instances
soap end(&soap); // clean up everything and close socket
}
}
soap done(&soap); // close master socket and detach environment
}
The soap serve dispatcher handles one request or multiple requests when HTTP keep-alive is enabled
(with the SOAP IO KEEPALIVE flag see Section 18.11).
The gSOAP functions that can be used are:
Function
soap new()
soap init(struct soap *soap)
soap bind(struct soap *soap, char *host, int port,
int backlog)
soap accept(struct soap *soap)
soap end(struct soap *soap)
soap free temp(struct soap *soap)
soap destroy(struct soap *soap)
soap done(struct soap *soap)
soap free(struct soap *soap)
40
Description
Allocates and Initializes gSOAP context
Initializes a stack-allocated gSOAP context (required once)
Returns master socket (backlog = max. queue size
for requests). When host==NULL: host is the machine on which the service runs
Returns slave socket
Clean up deserialized data (except class instances)
and temporary data
Clean up temporary data only
Clean up deserialized class instances (note: this
function will be renamed with option -n
Reset and detach context: close master/slave
sockets and remove callbacks
Detach and deallocate context (soap new())
The host name in soap bind may be NULL to indicate that the current host should be used.
The soap.accept timeout attribute of the gSOAP run-time environment specifies the timeout value for
a non-blocking soap accept(&soap) call. See Section 18.17 for more details on timeout management.
See Section 8.13 for more details on memory management.
A client application connects to this stand-alone service with the endpoint machine.cs.fsu.edu:18083.
A client may use the http:// prefix. When absent, no HTTP header is send and no HTTP-based
information will be communicated to the service.
7.2.4
How to Create a Multi-Threaded Stand-Alone Service
Multi-threading a Web Service is essential when the response times for handling requests by the
service are (potentially) long or when keep-alive is enabled, see Section 18.11. In case of long
response times, the latencies introduced by the unrelated requests may become prohibitive for a
successful deployment of a stand-alone service. When HTTP keep-alive is enabled, a client may
not close the socket on time, thereby preventing other clients from connecting.
gSOAP 2.0 and higher is thread safe and supports the implementation of multi-threaded stand-alone
services in which a thread is used to handle a request.
The following example illustrates the use of threads to improve the quality of service by handling
new requests in separate threads:
#include ”soapH.h”
#include <pthread.h>
#define BACKLOG (100) // Max. request backlog
int main(int argc, char **argv)
{
struct soap soap;
soap init(&soap);
if (argc < 2) // no args: assume this is a CGI application
{
soap serve(&soap); // serve request, one thread, CGI style
soap destroy(&soap); // dealloc C++ data
soap end(&soap); // dealloc data and clean up
}
else
{
soap.send timeout = 60; // 60 seconds
soap.recv timeout = 60; // 60 seconds
soap.accept timeout = 3600; // server stops after 1 hour of inactivity
soap.max keep alive = 100; // max keep-alive sequence
void *process request(void*);
struct soap *tsoap;
pthread t tid;
int port = atoi(argv[1]); // first command-line arg is port
SOAP SOCKET m, s;
m = soap bind(&soap, NULL, port, BACKLOG);
if (!soap valid socket(m))
exit(1);
41
fprintf(stderr, "Socket connection successful %d\n", m);
for (;;)
{
s = soap accept(&soap);
if (!soap valid socket(s))
{
if (soap.errnum)
{
soap print fault(&soap, stderr);
exit(1);
}
fprintf(stderr, "server timed out\n");
break;
}
fprintf(stderr, "Thread %d accepts socket %d connection from IP %d.%d.%d.%d\n",
i, s, (soap.ip>>24)&0xFF, (soap.ip>>16)&0xFF, (soap.ip>>8)&0xFF, soap.ip&0xFF);
tsoap = soap copy(&soap); // make a safe copy
if (!tsoap)
break;
pthread create(&tid, NULL, (void*(*)(void*))process request, (void*)tsoap);
}
}
soap done(&soap); // detach soap struct
return 0;
}
void *process request(void *soap)
{
pthread detach(pthread self());
soap serve((struct soap*)soap);
soap destroy((struct soap*)soap); // dealloc C++ data
soap end((struct soap*)soap); // dealloc data and clean up
soap done((struct soap*)soap); // detach soap struct
free(soap);
return NULL;
}
Note: the code does not wait for threads to join the main thread upon program termination.
The soap serve dispatcher handles one request or multiple requests when HTTP keep-alive is set
with SOAP IO KEEPALIVE. The soap.max keep alive value can be set to the maximum keep-alive calls
allowed, which is important to avoid a client from holding a thread indefinitely. The send and receive
timeouts are set to avoid (intentionally) slow clients from holding a socket connection too long. The
accept timeout is used to let the server terminate automatically after a period of inactivity.
The following example uses a pool of servers to limit the machine’s resource utilization:
#include ”soapH.h”
#include <pthread.h>
#define BACKLOG (100) // Max. request backlog
#define MAX THR (10) // Max. threads to serve requests
int main(int argc, char **argv)
{
42
struct soap soap;
soap init(&soap);
if (argc < 2) // no args: assume this is a CGI application
{
soap serve(&soap); // serve request, one thread, CGI style
soap destroy(&soap); // dealloc C++ data
soap end(&soap); // dealloc data and clean up
}
else
{
struct soap *soap thr[MAX THR]; // each thread needs a runtime environment
pthread t tid[MAX THR];
int port = atoi(argv[1]); // first command-line arg is port
SOAP SOCKET m, s;
int i;
m = soap bind(&soap, NULL, port, BACKLOG);
if (!soap valid socket(m))
exit(1);
fprintf(stderr, "Socket connection successful %d\n", m);
for (i = 0; i < MAX THR; i++)
soap thr[i] = NULL;
for (;;)
{
for (i = 0; i < MAX THR; i++)
{
s = soap accept(&soap);
if (!soap valid socket(s))
{
if (soap.errnum)
{
soap print fault(&soap, stderr);
continue; // retry
}
else
{
fprintf(stderr, "Server timed out\n");
break;
}
}
fprintf(stderr, "Thread %d accepts socket %d connection from IP %d.%d.%d.%d\n",
i, s, (soap.ip>>24)&0xFF, (soap.ip>>16)&0xFF, (soap.ip>>8)&0xFF, soap.ip&0xFF);
if (!soap thr[i]) // first time around
{
soap thr[i] = soap copy(&soap);
if (!soap thr[i])
exit(1); // could not allocate
}
else// recycle soap environment
{
pthread join(tid[i], NULL);
fprintf(stderr, ”Thread %d completed\n”, i);
43
soap destroy(soap thr[i]); // deallocate C++ data of old thread
soap end(soap thr[i]); // deallocate data of old thread
}
soap thr[i]->socket = s; // new socket fd
pthread create(&tid[i], NULL, (void*(*)(void*))soap serve, (void*)soap thr[i]);
}
}
for (i = 0; i < MAX THR; i++)
if (soap thr[i])
{
soap done(soap thr[i]); // detach context
free(soap thr[i]); // free up
}
}
return 0;
}
The following functions can be used to setup a gSOAP runtime environment (struct soap):
Function
soap init(struct soap *soap)
struct soap *soap new()
struct soap *soap copy(struct soap *soap)
soap done(struct soap *soap)
Description
Initializes a runtime environment (required only once)
Allocates, initializes, and returns a pointer to a runtime
environment
Allocates a new runtime environment and copies contents
of the argument environment such that the new environment does not share data with the argument environment
Reset, close communications, and remove callbacks
A new environment is initiated for each thread to guarantee exclusive access to runtime environments.
For clean termination of the server, the master socket can be closed and callbacks removed with
soap done(struct soap *soap).
The advantage of the code shown above is that the machine cannot be overloaded with requests,
since the number of active services is limited. However, threads are still started and terminated.
This overhead can be eliminated using a queue of requests (open sockets) as is shown in the code
below.
#include ”soapH.h”
#include <pthread.h>
#define BACKLOG (100) // Max. request backlog
#define MAX THR (10) // Size of thread pool
#define MAX QUEUE (1000) // Max. size of request queue
SOAP SOCKET queue[MAX QUEUE]; // The global request queue of sockets
int head = 0, tail = 0; // Queue head and tail
void *process queue(void*);
int enqueue(SOAP SOCKET);
SOAP SOCKET dequeue();
pthread mutex t queue cs;
pthread cond t queue cv;
44
int main(int argc, char **argv)
{
struct soap soap;
soap init(&soap);
if (argc < 2) // no args: assume this is a CGI application
{
soap serve(&soap); // serve request, one thread, CGI style
soap destroy(&soap); // dealloc C++ data
soap end(&soap); // dealloc data and clean up
}
else
{
struct soap *soap thr[MAX THR]; // each thread needs a runtime environment
pthread t tid[MAX THR];
int port = atoi(argv[1]); // first command-line arg is port
SOAP SOCKET m, s;
int i;
m = soap bind(&soap, NULL, port, BACKLOG);
if (!soap valid socket(m))
exit(1);
fprintf(stderr, "Socket connection successful %d\n", m);
pthread mutex init(&queue cs, NULL);
pthread cond init(&queue cv, NULL);
for (i = 0; i < MAX THR; i++)
{
soap thr[i] = soap copy(&soap);
fprintf(stderr, "Starting thread %d\n", i);
pthread create(&tid[i], NULL, (void*(*)(void*))process queue, (void*)soap thr[i]);
}
for (;;)
{
s = soap accept(&soap);
if (!soap valid socket(s))
{
if (soap.errnum)
{
soap print fault(&soap, stderr);
continue; // retry
}
else
{
fprintf(stderr, "Server timed out\n");
break;
}
}
fprintf(stderr, "Thread %d accepts socket %d connection from IP %d.%d.%d.%d\n",
i, s, (soap.ip>>24)&0xFF, (soap.ip>>16)&0xFF, (soap.ip>>8)&0xFF, soap.ip&0xFF);
while (enqueue(s) == SOAP EOM)
sleep(1);
}
for (i = 0; i < MAX THR; i++)
45
{
while (enqueue(SOAP INVALID SOCKET) == SOAP EOM)
sleep(1);
}
for (i = 0; i < MAX THR; i++)
{
fprintf(stderr, "Waiting for thread %d to terminate... ", i);
pthread join(tid[i], NULL);
fprintf(stderr, "terminated\n");
soap done(soap thr[i]);
free(soap thr[i]);
}
pthread mutex destroy(&queue cs);
pthread cond destroy(&queue cv);
}
soap done(&soap);
return 0;
}
void *process queue(void *soap)
{
struct soap *tsoap = (struct soap*)soap;
for (;;)
{
tsoap->socket = dequeue();
if (!soap valid socket(tsoap->socket))
break;
soap serve(tsoap);
soap destroy(tsoap);
soap end(tsoap);
fprintf(stderr, "served\n");
}
return NULL;
}
int enqueue(SOAP SOCKET sock)
{
int status = SOAP OK;
int next;
pthread mutex lock(&queue cs);
next = tail + 1;
if (next >= MAX QUEUE)
next = 0;
if (next == head)
status = SOAP EOM;
else
{
queue[tail] = sock;
tail = next;
}
pthread cond signal(&queue cv);
pthread mutex unlock(&queue cs);
return status;
46
}
SOAP SOCKET dequeue()
{
SOAP SOCKET sock;
pthread mutex lock(&queue cs);
while (head == tail)
pthread cond wait(&queue cv, &queue cs);
sock = queue[head++];
if (head >= MAX QUEUE)
head = 0;
pthread mutex unlock(&queue cs);
return sock;
}
Note: the plugin/threads.h and plugin/threads.c code can be used for a portable implementation.
Instead of POSIX calls, use MUTEX LOCK, MUTEX UNLOCK, and COND WAIT. These are wrappers
for Win API calls or POSIX calls.
7.2.5
How to Pass Application Data to Service Methods
The void *soap.user field can be used to pass application data to service methods. This field should
be set before the soap serve() call. The service method can access this field to use the applicationdependent data. The following example shows how a non-static database handle is initialized and
passed to the service methods:
{ ...
struct soap soap;
database handle type database handle;
soap init(&soap); soap.user = (void*)database handle;
...
soap serve(&soap); // call the remove method dispatcher to handle request
...
}
int ns myMethod(struct soap *soap, ...)
{ ...
fetch((database handle type*)soap->user);
// get data ...
return SOAP OK;
}
Another way to pass application data around in a more organized way is accomplished with plugins,
see Section 18.36.
7.2.6
Some Web Service Implementation Issues
The same client header file specification issues apply to the specification and implementation of a
SOAP Web service. Refer to
• 7.1.2 for namespace considerations.
47
• 7.1.5 for an explanation on how to change the encoding of the primitive types.
• 7.1.7 for a discussion on how the response element format can be controlled.
• 7.1.9 for details on how to pass multiple output parameters from a remote method.
• 7.1.11 for passing complex data types as output parameters.
• 7.1.13 for anonymizing the input and output parameter names.
7.2.7
How to Generate C++ Server Object Classes
Server object classes for C++ server applications are automatically generated by the gSOAP compiler.
There are two modes for generating classes. Use soapcpp2 option -i to generate improved class
definitions where the class’ member functions are the service methods. The older examples use a
C-like approach with globally defined service methods, which is illustated here with a calculator
example:
// Content of file "calc.h":
//gsoap ns service name: Calculator
//gsoap ns service style: rpc
//gsoap ns service encoding: encoded
//gsoap ns service location: http://www.cs.fsu.edu/˜engelen/calc.cgi
//gsoap ns schema namespace: urn:calc
//gsoap ns service method-action: add ””
int ns add(double a, double b, double &result); int ns sub(double a, double b, double &result);
int ns mul(double a, double b, double &result); int ns div(double a, double b, double &result);
The first three directives provide the service name which is used to name the service class, the
service location (endpoint), and the schema. The fourth directive defines the optional SOAPAction
for the method, which is a string associated with SOAP 1.1 operations. Compilation of this header
file with the gSOAP compiler soapcpp2 creates a new file soapCalculatorObject.h with the following
contents:
#include ”soapH.h”
class CalculatorObject : public soap
{ public:
Calculator() { ... };
˜Calculator() { ... }};
int serve() { return soap serve(this); };
};
This generated server object class can be included into a server application together with the
generated namespace table as shown in this example:
#include ”soapCalculatorObject.h” // get server object
#include ”Calculator.nsmap” // get namespace bindings
int main()
48
{
CalculatorObject c;
return c.serve(); // calls soap serve to serve as CGI application (using stdin/out)
}
int ns add(double a, double b, double &result)
{
result = a + b;
return SOAP OK;
}
... sub(), mul(), and div() implementations ...
You can use soapcpp2 compiler option -n together with -p to create a local namespaces table to
avoid link conflict when you need to combine multiple tables and/or multiple servers, see also
Sections 8.1 and 18.34, and you can use a C++ code namespace to create a namespace qualified
server object class, see Section 18.33.
7.2.8
How to Generate WSDL Service Descriptions
The gSOAP stub and skeleton compiler soapcpp2 generates WSDL (Web Service Description Language) service descriptions and XML Schema files when processing a header file. The compiler
produces one WSDL file for a set of remote methods. The names of the function prototypes of the
remote methods must use the same namespace prefix and the namespace prefix is used to name the
WSDL file. If multiple namespace prefixes are used to define remote methods, multiple WSDL files
will be created and each file describes the set of remote methods belonging to a namespace prefix.
In addition to the generation of the ns.wsdl file, a file with a namespace mapping table is generated
by the gSOAP compiler. An example mapping table is shown below:
struct Namespace namespaces[] =
{
{”SOAP-ENV”, ”http://schemas.xmlsoap.org/soap/envelope/”},
{”SOAP-ENC”, ”http://schemas.xmlsoap.org/soap/encoding/”},
¨
{”xsi”, ”http://www.w3.org/2001/XMLSchema-instance”, http://www.w3.org/*/XMLSchemainstance”},
¨
{”xsd”, ”http://www.w3.org/2001/XMLSchema”, http://www.w3.org/*/XMLSchema”},
{”ns”, ”http://tempuri.org”},
{NULL, NULL}
};
This file can be incorporated in the client/service application, see Section 9.4 for details on namespace mapping tables.
To deploy a Web service, copy the compiled CGI service application to the designated CGI directory of your Web server. Make sure the proper file permissions are set (chmod 755 calc.cgi for
Unix/Linux). You can then publish the WSDL file on the Web by placing it in the appropriate
Web server directory.
The gSOAP compiler also generates XML Schema files for all C/C++ complex types (e.g. structs
and classes) when declared with a namespace prefix. These files are named ns.xsd, where ns is the
49
namespace prefix used in the declaration of the complex type. The XML Schema files do not have
to be published as the WSDL file already contains the appropriate XML Schema definitions.
7.2.9
Example
For example, suppose the following methods are defined in the header file:
typedef double
int ns add(xsd
int ns sub(xsd
int ns sqrt(xsd
xsd double;
double a, xsd double b, xsd double &result);
double a, xsd double b, xsd double &result);
double a, xsd double &result);
Then, one WSDL file will be created with the file name ns.wsdl that describes all three remote
methods:
<?xml version="1.0" encoding="UTF-8"?>
<definitions name="Service"
xmlns="http://schemas.xmlsoap.org/wsdl/"
targetNamespace="http://location/Service.wsdl"
xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:SOAP-ENC="http://schemas.xmlsoap.org/soap/encoding/"
xmlns:SOAP="http://schemas.xmlsoap.org/wsdl/soap/"
xmlns:WSDL="http://schemas.xmlsoap.org/wsdl/"
xmlns:xsd="http://www.w3.org/2000/10/XMLSchema"
xmlns:tns="http://location/Service.wsdl"
xmlns:ns="http://tempuri.org">
<types>
<schema
xmlns="http://www.w3.org/2000/10/XMLSchema"
targetNamespace="http://tempuri.org"
xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:SOAP-ENC="http://schemas.xmlsoap.org/soap/encoding/">
<complexType name="addResponse">
<all>
<element name="result" type="double" minOccurs="0" maxOccurs="1"/>
</all>
<anyAttribute namespace="##other"/>
</complexType>
<complexType name="subResponse">
<all>
<element name="result" type="double" minOccurs="0" maxOccurs="1"/>
</all>
<anyAttribute namespace="##other"/>
</complexType>
<complexType name="sqrtResponse">
<all>
<element name="result" type="double" minOccurs="0" maxOccurs="1"/>
</all>
<anyAttribute namespace="##other"/>
50
</complexType>
</schema>
</types>
<message name="addRequest">
<part name="a" type="xsd:double"/>
<part name="b" type="xsd:double"/>
</message>
<message name="addResponse">
<part name="result" type="xsd:double"/>
</message>
<message name="subRequest">
<part name="a" type="xsd:double"/>
<part name="b" type="xsd:double"/>
</message>
<message name="subResponse">
<part name="result" type="xsd:double"/>
</message>
<message name="sqrtRequest">
<part name="a" type="xsd:double"/>
</message>
<message name="sqrtResponse">
<part name="result" type="xsd:double"/>
</message>
<portType name="ServicePortType">
<operation name="add">
<input message="tns:addRequest"/>
<output message="tns:addResponse"/>
</operation>
<operation name="sub">
<input message="tns:subRequest"/>
<output message="tns:subResponse"/>
</operation>
<operation name="sqrt">
<input message="tns:sqrtRequest"/>
<output message="tns:sqrtResponse"/>
</operation>
</portType>
<binding name="ServiceBinding" type="tns:ServicePortType">
<SOAP:binding style="rpc" transport="http://schemas.xmlsoap.org/soap/http"/>
<operation name="add">
<SOAP:operation soapAction="http://tempuri.org#add"/>
<input>
<SOAP:body use="encoded" namespace="http://tempuri.org"
encodingStyle="http://schemas.xmlsoap.org/soap/encoding/"/>
</input>
<output>
<SOAP:body use="encoded" namespace="http://tempuri.org"
encodingStyle="http://schemas.xmlsoap.org/soap/encoding/"/>
</output>
</operation>
<operation name="sub">
51
<SOAP:operation soapAction="http://tempuri.org#sub"/>
<input>
<SOAP:body use="encoded" namespace="http://tempuri.org"
encodingStyle="http://schemas.xmlsoap.org/soap/encoding/"/>
</input>
<output>
<SOAP:body use="encoded" namespace="http://tempuri.org"
encodingStyle="http://schemas.xmlsoap.org/soap/encoding/"/>
</output>
</operation>
<operation name="sqrt">
<SOAP:operation soapAction="http://tempuri.org#sqrt"/>
<input>
<SOAP:body use="encoded" namespace="http://tempuri.org"
encodingStyle="http://schemas.xmlsoap.org/soap/encoding/"/>
</input>
<output>
<SOAP:body use="encoded" namespace="http://tempuri.org"
encodingStyle="http://schemas.xmlsoap.org/soap/encoding/"/>
</output>
</operation>
</binding>
<service name="Service">
<port name="ServicePort" binding="tns:ServiceBinding">
<SOAP:address location="http://location/Service.cgi"/>
</port>
</service>
</definitions>
7.2.10
How to Parse and Import WSDL Service Descriptions to Develop Clients and
Servers
Note: see README.txt in the wsdl directory for installation instructions for the WSDL parser and
importer.
The creation of SOAP Web Services applications from a WSDL service description is a two-step
process.
First, execute wsdl2h file.wsdl which generates the a C++ header file file.h (use wsdl2h -c file.wsdl
to generate pure C code). You can provide a URL instead of a file name, when applicable. The
generated header file is a Web service specification that contains the parameter types and service
function definitions. The functions are represented as function prototypes. The file contains various
annotations related to the Web service. The header file must be processed by the gSOAP compiler.
You cannot use it with a C/C++ compiler directly.
Second, the header file file.h is processed by the gSOAP compiler by executing soapcpp2 -i file.h. This
creates the C/C++ source files to build a client application, see 7.1. In addition, this generates a
client proxy object declared in soapServiceProxy.h, where Service is the name of the service defined
in the WSDL. To use this object, include the soapServiceProxy.h and Service.nsmap files in your C++
client application. The Service class provides the remote Web service methods as class members.
52
Consider the following example commands (entered at the command prompt):
$ wsdl2h -o XMethodsQuery.h http://www.xmethods.net/wsdl/query.wsdl
...
$ soapcpp2 -i XMethodsQuery.h
The first command generates XMethodsQuery.h from the WSDL at the specified URL. The header
file is then processed by the gSOAP compiler to generate the stubs and skeletons. See XMethodsQuery.h for the types and service functions. A C++ client application may use the generated
soapXMethodsQuerySoapProxy.h and soapXMethodsQuerySoapProxy.cpp class and XMethodsQuerySoap.nsmap
XML namespace table to access the Web service. Both need to be #include-d in your source. Then
compile and link the soapC.cpp and stdsoap2.cpp sources to complete the build.
When parsing a WSDL, the output file name is the WSDL input file name with extension .h
instead of .wsdl. When an input file is absent or a WSDL file from a Web location is accessed, the
header output will be produced on the standard output. Schema files (.xsd) can also be parsed and
processed.
The wsdl2h command-line options are:
Option
-a
-c
-d
-e
-f
-g
-h
-I path
-j
-l
-m
-n name
-N name
-o file
-p
-q name
-r host:port
-s
-t file
-u
-v
-w
-x
-y
-z
-?
Description
generate indexed struct names for local elements with anonymous types
generate C source code
use DOM to populate xs:any and xsd:anyType elements
don’t qualify enum names
This option is for backward compatibility with gSOAP 2.4.1 and earlier.
The option does not produce code that conforms to WS-I Basic Profile 1.0a.
generate flat C++ class hierarchy for schema extensions
generate global top-level element declarations
print help information
use path to find files
don’t generate SOAP ENV Header and SOAP ENV Detail definitions
include license information in output
use xsd.h module to import primitive types
use name as the base namespace prefix name instead of ns
use name as the base namespace prefix name for service namespaces
output to file
create polymorphic types with C++ inheritance hierarchy with base xsd anyType
This is automatically performed when WSDL contains polymorphic definitions
use name for the C++ namespace for all service declarations
connect via proxy host and port
don’t generate STL code (no std::string and no std::vector)
use type map file instead of the default file typemap.dat
don’t generate unions
verbose output
always wrap response parameters in a response struct
don’t generate XML any/anyAttribute extensibility elements
generate typedef synonyms for structs and enums
generate pointer-based arrays for backward compatibility ¡ gSOAP 2.7.6e
don’t generate USCORE (replace with UNICODE x005f)
print help information
53
7.2.11
The typemap.dat File
A typemap.dat file for wsdl2h contains custom XML Schema and C/C++ type bindings. An internal
table is used by default.
An example typemap file is:
# This file contains custom definitions of the XML Schema types and
# C/C++ types for your project, and XML namespace prefix definitions.
# The wsdl2h WSDL importer consults this file to determine bindings.
[
// This comment will be included in the generated .h file
// You can include any additional declarations, includes, imports, etc.
// within [ ] sections. The brackets MUST appear at the start of a line
]
# XML namespace prefix definitions can be provided to override the
# default choice of ns1, ns2, ... prefixes. For example:
i = "http://www.soapinterop.org/"
s = "http://www.soapinterop.org/xsd"
#
#
#
#
#
#
#
#
Type definitions are of the form: # type = declaration | use | pointer-use
where
type is the XML Schema type (or an application type in a namespace
that has a prefix definition given as above).
declaration is an optional C/C++ type declaration
use is how the type is referred to in code
pointer-use is how the type should be referred to as a pointer (opt)
Example XML Schema and C/C++ type bindings:
xsd int = | int
xsd string = | char* | char*
xsd boolean = enum xsd boolean false , true ; | enum xsd boolean
xsd base64Binary = class xsd base64Binary unsigned char * ptr; int size; ; |
xsd base64Binary | xsd base64Binary
# You can extend structs and classes with member data and functions.
# For example, adding a constructor to ns myClass: ns myClass = $ ns myClass();
# The general form is # class name = $ member;
The i and s prefixes are declared such that the header file output by the WSDL parser will use
these to produce C/C++ code. XML Schema types are associated with an optional C/C++ type
declaration, a use reference, and a pointer-use reference. The pointer-use reference of the xsd byte
type for example, is int* because char* is reserved for strings.
7.2.12
How to Use Client Functionalities Within a Service
A gSOAP service implemented with CGI may make direct client calls to other services from within
its service operations, without setting up a new context. A stand-alone service application must
54
setup a new soap struct context, e.g. using soap copy and delete it after the call.
The server-side client call is best illustrated with an example. The following example is a more
sophisticated example that combines the functionality of two Web services into one new SOAP Web
service. The service provides a currency-converted stock quote. To serve a request, the service in
turn requests the stock quote and the currency-exchange rate from two XMethods services.
In addition to being a client of two XMethods services, this service application can also be used as a
client of itself to test the implementation. As a client invoked from the command-line, it will return
a currency-converted stock quote by connecting to a copy of itself installed as a CGI application
on the Web to retrieve the quote after which it will print the quote on the terminal.
The header file input to the gSOAP compiler is given below:
// Contents of file ”quotex.h”:
int ns1 getQuote(char *symbol, float &result); // XMethods delayed stock quote service remote
method
int ns2 getRate(char *country1, char *country2, float &result); // XMethods currency-exchange
service remote method
int ns3 getQuote(char *symbol, char *country, float &result); // the new currency-converted
stock quote service
The quotex.cpp client/service application source is:
// Contents of file ”quotex.cpp”:
#include ”soapH.h” // include generated proxy and SOAP support
int main(int argc, char **argv)
{
struct soap soap;
float q;
soap init(&soap);
if (argc <= 2)
soap serve(&soap);
else if (soap call ns3 getQuote(&soap, "http://www.cs.fsu.edu/~engelen/quotex.cgi",
"", argv[1], argv[2], q))
soap print fault(&soap, stderr);
else
printf("\nCompany %s: %f (%s)\n", argv[1], q, argv[2]);
return 0;
}
int ns3 getQuote(struct soap *soap, char *symbol, char *country, float &result)
{
float q, r;
int socket = soap->socket; // save socket (stand-alone service only, does not support keep-alive)
if (soap call ns1 getQuote(soap, "http://services.xmethods.net/soap", "", symbol, &q)
== 0 &&
soap call ns2 getRate(soap, "http://services.xmethods.net/soap", NULL, "us", country, &r) == 0)
{
result = q*r;
soap->socket = socket;
55
return SOAP OK;
}
soap->socket = socket;
return SOAP FAULT; // pass soap fault messages on to the client of this app
}
/* Since this app is a combined client-server, it is put together with
one header file that describes all remote methods. However, as a consequence we
have to implement the methods that are not ours. Since these implementations are
never called (this code is client-side), we can make them dummies as below.
/
int ns1 getQuote(struct soap *soap, char *symbol, float &result)
{ return SOAP NO METHOD; } // dummy: will never be called
int ns2 getRate(struct soap *soap, char *country1, char *country2, float &result)
{ return SOAP NO METHOD; } // dummy: will never be called
struct Namespace namespaces[] =
{
{”SOAP-ENV”, ”http://schemas.xmlsoap.org/soap/envelope/”},
{”SOAP-ENC”, ”http://schemas.xmlsoap.org/soap/encoding/”},
{”xsi”, ”http://www.w3.org/2001/XMLSchema-instance”, ”http://www.w3.org/*/XMLSchemainstance”},
{”xsd”, ”http://www.w3.org/2001/XMLSchema”, ”http://www.w3.org/*/XMLSchema”},
{”ns1”, ”urn:xmethods-delayed-quotes”},
{”ns2”, ”urn:xmethods-CurrencyExchange”},
{”ns3”, ”urn:quotex”},
{NULL, NULL}
};
To compile:
soapcpp2 quotex.h
g++ -o quotex.cgi quotex.cpp soapC.cpp soapClient.cpp soapServer.cpp stdsoap2.cpp -lsocket -lxnet
-lnsl
Note: under Linux and Mac OS X you can often omit the -l libraries.
The quotex.cgi executable is installed as a CGI application on the Web by copying it in the designated
directory specific to your Web server. After this, the executable can also serve to test the service.
For example
quotex.cgi IBM uk
returns the quote of IBM in uk pounds by communicating the request and response quote from
the CGI application. See http://xmethods.com/detail.html?id=5 for details on the currency
abbreviations.
When combining clients and service functionalities, it is required to use one header file input to
the compiler. As a consequence, however, stubs and skeletons are available for all remote methods,
while the client part will only use the stubs and the service part will use the skeletons. Thus,
dummy implementations of the unused remote methods need to be given which are never called.
56
Three WSDL files are created by gSOAP: ns1.wsdl, ns2.wsdl, and ns3.wsdl. Only the ns3.wsdl file
is required to be published as it contains the description of the combined service, while the others
are generated as a side-effect (and in case you want to develop these separate services).
7.3
How to Use gSOAP for Asynchronous One-Way Message Passing
SOAP RPC client-server interaction is synchronous: the client blocks until the server responds to
the request. gSOAP also supports asynchronous one-way message passing and the interoperable
synchronous one-way message passing over HTTP. The two styles are similar, but only the latter is
interoperable and is compliant to Basic Profile 1.0. The interoperable synchronous one-way message
passing style over HTTP is discussed in Section 7.4 below.
SOAP messaging routines are declared as function prototypes, just like remote methods for SOAP
RPC. However, the output parameter is a void type to indicate the absence of a return value.
For example, the following header file specifies a event message for SOAP messaging:
int ns event(int eventNo, void dummy);
The gSOAP stub and skeleton compiler generates the following functions in soapClient.cpp:
int soap send ns event(struct soap *soap, const char URL, const char action, int event);
int soap recv ns event(struct soap *soap, struct ns event *dummy);
The soap send ns event function transmits the message to the destination URL by opening a socket
and sending the SOAP encoded message. The socket will remain open after the send and has to
be closed with soap closesock(). The open socket connection can also be used to obtain a service
response, e.g. with a soap recv function call.
The soap recv ns event function waits for a SOAP message on the currently open socket (soap.socket)
and fills the struct ns event with the ns event parameters (e.g. int eventNo). The struct ns event is
automatically created by gSOAP and is a mirror image of the ns event parameters:
struct ns event
{ int eventNo;
}
The gSOAP generated soapServer.cpp code includes a skeleton routine to accept the message. (The
skeleton routine does not respond with a SOAP response message.)
int soap serve ns event(struct soap *soap);
The skeleton routine calls the user-implemented ns event(struct soap *soap, int eventNo) routine (note
the absence of the void parameter!).
As usual, the skeleton will be automatically called by the remote method request dispatcher that
handles both the remote method requests (RPCs) and messages:
57
main()
{ soap serve(soap new());
}
int ns event(struct soap *soap, int eventNo)
{
... // handle event
return SOAP OK;
}
7.4
One-Way Message Passing over HTTP
One-way SOAP message passing over HTTP as defined by the SOAP specification and Basic Profile
1.0 is synchrounous, meaning that the server must respond with an HTTP OK header and an empty
body. To implement synchrounous one-way messaging, the same setup for asynchrounous one-way
messaing discussed in Section 7.3 is used, but with one simple addition at the client and server
side.
At the server side, we need to return an empty HTTP OK response. This is accomplished as
follows. For each one-way operation implemented in C/C++, we replace the return SOAP OK with:
int ns event(struct soap *soap, int eventNo)
{
... // handle event
return soap send empty response(soap, SOAP OK); // SOAP OK: return HTTP 202 ACCEPTED
}
At the client side, the empty response header must be parsed as follows:
if (soap send ns event(soap, eventNo) != SOAP OK
|| soap recv empty response(soap) != SOAP OK)
soap print fault(soap, stderr);
...
The synchronous (and asynchronous) one-way messaging supports HTTP keep-alive and chunking.
7.5
How to Use the SOAP Serializers and Deserializers to Save and Load Application Data
The gSOAP stub and skeleton compiler generates serializers and deserializers for all user-defined
data structures that are specified in the header file input to the compiler. The serializers and
deserializers can be found in the generated soapC.cpp file. These serializers and deserializers can
be used separately by an application without the need to build a full client or service application.
This is useful for applications that need to save or export their data in XML or need to import or
load data stored in XML format.
The following attributes can be set to control the destination and source for serialization and
deserialization:
58
Variable
int soap.socket
ostream *soap.os
istream *soap.is
int soap.sendfd
int soap.recvfd
Description
socket file descriptor for input and output or -1
(C++ only) output stream used for send operations
(C++ only) input stream used for receive operations
when soap socket<0, this fd is used for send operations
when soap socket<0, this fd is used for receive operations
The following initializing and finalizing functions can be used:
Function
void soap begin send(struct soap*)
int soap end send(struct soap*)
int soap begin recv(struct soap*)
int soap end recv(struct soap*)
Description
start a send/write phase
flush the buffer
start a rec/read phase (if an HTTP header is present, parse it first)
perform a id/href consistency check on deserialized data
These operations do not open or close the connections. The application should open and close
connections or files and set the soap.socket, soap.os or soap.sendfd, soap.is or soap.recvfd streams or
descriptors. When soap.socket<0 and none of the streams and descriptors are set, then the standard
input and output will be used.
See also Section 8.12 to control the I/O buffering and content encoding such as compression and
DIME encoding.
7.5.1
Serializing a Data Type
To serialize a data type to a stream, two functions should be called to prepare for serialization of
the data and to send the data, respectively. The first function, soap serialize, analyzes pointers and
determines if multi-references are required to encode the data and if cycles are present the object
graph. The second function, soap put, produces the XML output on a stream.
The soap serialize and soap put function names are specific to a data type. For example, soap serialize float(&soap,
&d) is called to serialize an float value and soap put float(&soap, &d, ”number”, NULL) is called to output the floating point value in SOAP tagged with the name <number>. To initialize data, the
soap default function of a data type can be used. For example, soap default float(&soap, &d) initializes
the float to 0.0. The soap default functions are useful to initialize complex data types such as arrays,
structs, and class instances. Note that the soap default functions do not need the gSOAP runtime
environment as a first parameter.
The following table lists the type naming conventions used by gSOAP:
59
Type
char*
wchar t*
char
bool
double
int
float
long
LONG64
long long
short
time t
unsigned char
unsigned int
unsigned long
ULONG64
unsigned long long
unsigned short
T[N]
T*
struct Name
class Name
enum Name
Type Name
string
wstring
byte
bool
double
int
float
long
LONG64 (Win32)
LONG64 (Unix/Linux)
short
time
unsignedByte
unsignedInt
unsignedLong
unsignedLONG64 (Win32)
unsignedLONG64 (Unix/Linux)
unsignedShort
ArrayNOfType where Type is the type name of T
PointerToType where Type is the type name of T
Name
Name
Name
Consider for example the following C code with a declaration of p as a pointer to a struct ns Person:
struct ns Person { char *name; } *p;
To serialize p, its address is passed to the function soap serialize PointerTons Person generated for this
type by the gSOAP compiler:
soap serialize PointerTons Person(&soap, &p);
The address of p is passed, so the serializer can determine whether p was already serialized and to
discover cycles in graph data structures. To generate the output, the address of p is passed to the
function soap put PointerTons Person together with the name of an XML element and an optional
type string (to omit a type, use NULL):
soap begin send(&soap);
soap put PointerTons Person(&soap, &p, ”ns:element-name”, ”ns:type-name”);
soap end send(&soap);
This produces:
<ns:element-name xmlns:SOAP-ENV="..." xmlns:SOAP-ENC="..." xmlns:ns="..."
... xsi:type="ns:type-name">
<name xsi:type="xsd:string">...</name>
</ns:element-name>
60
The serializer is initialized with the soap begin send(soap) function and closed with soap end send(soap).
All temporary data structures and data structures deserialized on the heap are destroyed with the
soap destroy and soap end functions (in this order).
The soap done function should be used to reset the context, i.e. the last use of the context. To
detach and deallocate the context, use soap free.
To remove the temporary data only and keep the deserialized data on the heap, use soap free temp.
Temporary data structures are only created if the encoded data uses pointers. Each pointer in
the encoded data has an internal hash table entry to determine all multi-reference parts and cyclic
parts of the complete data structure.
You can assign an output stream to soap.os or a file descriptor to soap.sendfd. For example
soap.sendfd = open(file, O RDWR|O CREAT, S IWUSR|S IRUSR);
soap serialize PointerTons Person(&soap, &p);
soap begin send(&soap);
soap put PointerTons Person(&soap, &p, ”ns:element-name”, ”ns:type-name”);
soap end send(&soap);
The soap serialize function is optional. It MUST be used when the object graph contains cycles. It
MUST be called to preserved the logical coherence of pointer-based data structures, where pointers
may refer to co-referenced objects. By calling soap serialize, data structures shared through pointers
are serialized only once and referenced in XML using id-refs attributes. This actual id-refs used
depend on the SOAP encoding. To turn off SOAP encoding, remove or avoid using the SOAP-ENV
and SOAP-ENC namespace bindings in the namespace table. In addition, the SOAP XML TREE and
SOAP XML GRAPH flags can be used to control the output.
To save the data as an XML tree (with one root) without any id-ref attributes, use the SOAP XML TREE
flag. The data structure MUST NOT contain pointer-based cycles.
To preserve the exact structure of the data object graph and create XML with one root, use the
SOAP XML GRAPH output-mode flag (see Section 8.12). Use this flag and the soap serialize function
to prepare the serialization of data with in-line id-ref attributes. Using the SOAP XML GRAPH flag
assures the preservation of the logical structure of the data
For example, to encode the contents of two variables var1 and var2 that may share data throug
pointer structures, the serializers are called before the output routines:
T1 var1;
T2 var2;
struct soap soap;
...
soap init(&soap); // initialize
[soap omode(&soap, flags);] // set output-mode flags (e.g. SOAP ENC XML|SOAP ENC ZLIB)
soap begin(&soap); // start new (de)serialization phase
soap set omode(&soap, SOAP XML GRAPH);
soap serialize Type1(&soap, &var1);
soap serialize Type2(&soap, &var2);
...
[soap.socket = a socket file descriptor;] // when using sockets
[soap.os = an output stream;] // C++
61
[soap.sendfd = an output file descriptor;] // C
soap
soap
soap
...
soap
soap
soap
soap
...
begin send(&soap);
put Type1(&soap, &var1, ”[namespace-prefix:]element-name1”, ”[namespace-prefix:]type-name1”);
put Type2(&soap, &var2, ”[namespace-prefix:]element-name2”, ”[namespace-prefix:]type-name2”);
end send(&soap); // flush
destroy(&soap); // remove deserialized C++ objects
end(&soap); // remove deserialized data structures
done(&soap); // finalize last use of this environment
where Type1 is the type name of T1 and Type2 is the type name of T2 (see table above). The
strings [namespace-prefix:]type-name1 and [namespace-prefix:]type-name2 describe the schema types of the
elements. Use NULL to omit this type information.
For serializing class instances, method invocations MUST be used instead of function calls, for
example obj.soap serialize(&soap) and obj.soap put(&soap, ”elt”, ”type”). This ensures that the proper
serializers are used for serializing instances of derived classes.
You can serialize a class instance to a stream as follows:
struct soap soap;
myClass obj;
soap init(&soap); // initialize
soap begin(&soap); // start new (de)serialization phase
soap set omode(&soap, SOAP XML GRAPH);
obj.serialize(&soap);
soap.os = cout; // send to cout
soap begin send(&soap);
obj.put(&soap, ”[namespace-prefix:]element-name1”, ”[namespace-prefix:]type-name1”);
...
soap end send(&soap); // flush
soap destroy(&soap); // remove deserialized C++ objects
soap end(&soap); // remove deserialized data
soap done(&soap); // finalize last use of this environment
When you declare a soap struct pointer as a data member in a class, you can overload the <<
operator to serialize the class to streams:
ostream &operator<<(ostream &o, const myClass &e)
{
if (!e.soap)
... error: need a soap struct to serialize (could use global struct) ...
else
{
ostream *os = e.soap->os;
e.soap->os = &o;
soap set omode(e.soap, SOAP XML GRAPH);
e.serialize(e.soap);
soap begin send(e.soap);
e.put(e.soap, ”myClass”, NULL);
soap end send(e.soap);
62
e.soap->os = os;
soap clr omode(e.soap, SOAP XML GRAPH);
}
return o;
}
Of course, when you construct an instance you must set its soap struct to a valid environment.
Deserialized class instances with a soap struct data member will have their soap structs set automatically, see Section 8.13.2.
In principle, XML output for a data structure can be produced with soap put without calling the
soap serialize function first. In this case, the result is similar to SOAP XML TREE which means that
no id-refs are output. Cycles in the data structure will crash the serialization algorithm, even when
the SOAP XML GRAPH is set.
Consider the following struct:
// Contents of file ”tricky.h”:
struct Tricky
{
int *p;
int n;
int *q;
};
The following fragment initializes the pointer fields p and q to the value of field n:
struct soap soap;
struct Tricky X;
X.n = 1;
X.p = &X.n;
X.q = &X.n;
soap init(&soap);
soap begin(&soap);
soap serialize Tricky(&soap, &X);
soap put Tricky(&soap, &X, "Tricky", NULL);
soap end(&soap); // Clean up temporary data used by the serializer
What is special about this data structure is that n is ’fixed’ in the Tricky structure, and p and q
both point to n. The gSOAP serializers strategically place the id-ref attributes such that n will be
identified as the primary data source, while p and q are serialized with ref/href attributes.
The resulting output is:
<Tricky xsi:type="Tricky">
<p href="#2"/> <n xsi:type="int">1</n> <q href="#2"/> <r xsi:type="int">2</r> </Tricky>
<id id="2" xsi:type="int">1</id>
which uses an independent element at the end to represent the multi-referenced integer, assuming
the SOAP-ENV and SOAP-ENC namespaces indicate SOAP 1.1 encoding.
With the SOAP XML GRAPH flag the output is:
63
<Tricky xsi:type="Tricky">
<p href="#2"/> <n id="2" xsi:type="int">1</n> <q href="#2"/> </Tricky>
In this case, the XML is self-contained and multi-referenced data is accurately serialized. The
gSOAP generated deserializer for this data type will be able to accurately reconstruct the data
from the XML (on the heap).
7.5.2
Deserializing a Data Type
To deserialize a data type, its soap get function is used. The outline of a program that deserializes
two variables var1 and var2 is for example:
T1 var1;
T2 var2;
struct soap soap;
...
soap init(&soap); // initialize at least once
[soap imode(&soap, flags);] // set input-mode flags
soap begin(&soap); // begin new decoding phase
[soap.is = an input stream;] // C++
[soap.recvfd = an input file desriptpr;] // C
soap begin recv(&soap); // if HTTP/MIME/DIME/GZIP headers are present, parse them
if (!soap get Type1(&soap, &var1, ”[namespace-prefix:]element-name1”, ”[namespace-prefix:]typename1”))
... error ...
if (!soap get Type2(&soap, &var2, ”[namespace-prefix:]element-name2”, ”[namespace-prefix:]typename1”))
... error ...
...
soap end recv(&soap); // check consistency of id/hrefs
soap destroy(&soap); // remove deserialized C++ objects
soap end(&soap); // remove deserialized data
soap done(&soap); // finalize last use of the environment
The strings [namespace-prefix:]type-name1 and [namespace-prefix:]type-name2 are the schema types of the
elements and should match the xsi:type attribute of the receiving message. To omit the match,
use NULL as the type. For class instances, method invocation can be used instead of a function call
if the object is already instantiated, i.e. obj.soap get(&soap, ”...”, ”...”).
The soap begin call resets the deserializers. The soap destroy and soap end calls remove the temporary
data structures and the decoded data that was placed on the heap.
To remove temporary data while retaining the deserialized data on the heap, the function soap free temp
should be called instead of soap destroy and soap end.
One call to the soap get Type function of a type Type scans the entire input to process its XML
content and to capture SOAP 1.1 independent elements (which contain multi-referenced objects).
As a result, soap.error will set to SOAP EOF. Also storing multiple objects into one file will fail to
decode them properly with multiple soap get calls. A well-formed XML document should only have
one root anyway, so don’t save multiple objects into one file. If you must save multiple objects,
64
create a linked list or an array of objects and save the linked list or array. You could use the
soap in Type function instead of the soap get Type function. The soap in Type function parses one
XML element at a time.
You can deserialize class instances from a stream as follows:
myClass obj;
struct soap soap;
soap init(&soap); // initialize
soap begin(&soap); // begin new decoding phase
soap.is = cin; // read from cin
soap begin recv(&soap); // if HTTP header is present, parse it
if (!obj.get(&soap, ”myClass”, NULL)
... error ...
soap end recv(&soap); // check consistency of id/hrefs
...
soap destroy(&soap); // remove deserialized C++ objects
soap end(&soap); // remove deserialized data
soap done(&soap); // finalize last use of the environment
When you declare a soap struct pointer as a data member in a class, you can overload the >>
operator to parse and deserialize a class instance from a stream:
istream &operator>>(istream &i, myClass &e)
{
if (!e.soap)
... error: need soap struct to deserialize (could use global struct)...
istream *is = e.soap->is;
e.soap->is = &i;
if (soap begin recv(e.soap) || e.in(e.soap, NULL, NULL) || soap end recv(e.soap))
... error ...
e.soap->is = is;
return i;
}
7.5.3
Example
As an example, consider the following data type declarations:
// Contents of file ”person.h”:
typedef char *xsd string;
typedef char *xsd Name;
typedef unsigned int xsd unsignedInt;
enum ns Gender {male, female};
class ns Address
{
public:
xsd string street;
xsd unsignedInt number;
xsd string city;
65
};
class ns Person
{
public:
xsd Name name;
enum ns Gender gender;
ns Address address;
ns Person *mother;
ns Person *father;
};
The following program uses these data types to write to standard output a data structure that
contains the data of a person named ”John” living at Downing st. 10 in Londen. He has a mother
”Mary” and a father ”Stuart”. After initialization, the class instance for ”John” is serialized and
encoded in XML to the standard output stream using gzip compression (requires the Zlib library,
compile sources with -DWITH GZIP):
// Contents of file ”person.cpp”:
#include ”soapH.h”
int main()
{
struct soap soap;
ns Person mother, father, john;
mother.name = "Mary";
mother.gender = female;
mother.address.street = "Downing st.";
mother.address.number = 10;
mother.address.city = "London";
mother.mother = NULL;
mother.father = NULL;
father.name = "Stuart";
father.gender = male;
father.address.street = "Main st.";
father.address.number = 5;
father.address.city = "London";
father.mother = NULL;
father.father = NULL;
john.name = "John";
john.gender = male;
john.address = mother.address;
john.mother = &mother;
john.father = &father;
soap init(&soap);
soap omode(&soap, SOAP ENC ZLIB|SOAP XML GRAPH); // see 8.12
soap begin(&soap);
soap begin send(&soap);
john.soap serialize(&soap);
john.soap put(&soap, "johnnie", NULL);
soap end send(&soap);
soap destroy(&soap);
soap end(&soap);
66
soap done(&soap);
}
struct Namespace namespaces[] =
{
{”SOAP-ENV”, ”http://schemas.xmlsoap.org/soap/envelope/”},
{”SOAP-ENC”,”http://schemas.xmlsoap.org/soap/encoding/”},
{”xsi”, ”http://www.w3.org/2001/XMLSchema-instance”},
{”xsd”, ”http://www.w3.org/2001/XMLSchema”},
{”ns”, ”urn:person”}, // Namespace URI of the “Person” data type
{NULL, NULL}
};
The header file is processed and the application compiled on Linux/Unix with:
soapcpp2 person.h
g++ -DWITH GZIP -o person person.cpp soapC.cpp stdsoap2.cpp -lsocket -lxnet -lnsl -lz
(Depending on your system configuration, the libraries libsocket.a, libxnet.a, libnsl.a are required.
Compiling on Linux typically does not require the inclusion of those libraries.) See 18.25 for details
on compression with gSOAP.
Running the person application results in the compressed XML output:
<johnnie xsi:type="ns:Person" xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:SOAP-ENC="http://schemas.xmlsoap.org/soap/encoding/"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:xsd="http://www.w3.org/2001/XMLSchema"
xmlns:ns="urn:person"
SOAP-ENV:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/">
<name xsi:type="xsd:Name">John</name>
<gender xsi:type="ns:Gender">male</gender>
<address xsi:type="ns:Address">
<street id="3" xsi:type="xsd:string">Dowling st.</street>
<number xsi:type="unsignedInt">10</number>
<city id="4" xsi:type="xsd:string">London</city>
</address>
<mother xsi:type="ns:Person">
<name xsi:type="xsd:Name">Mary</name>
<gender xsi:type="ns:Gender">female</gender>
<address xsi:type="ns:Address">
<street href="#3"/>
<number xsi:type="unsignedInt">5</number>
<city href="#4"/>
</address>
</mother>
<father xsi:type="ns:Person">
<name xsi:type="xsd:Name">Stuart</name>
<gender xsi:type="ns:Gender">male</gender>
<address xsi:type="ns:Address">
<street xsi:type="xsd:string">Main st.</street>
<number xsi:type="unsignedInt">13</number>
67
<city href="#4"/>
</address>
</father>
</johnnie>
The following program fragment decodes this content from standard input and reconstructs the
original data structure on the heap:
#include ”soapH.h”
int main()
{
struct soap soap;
ns Person *mother, *father, *john = NULL;
soap init(&soap);
soap imode(&soap, SOAP ENC ZLIB); // optional: gzip is detected automatically
soap begin(&soap);
soap begin recv(&soap);
if (soap get ns Person(&soap, john, ”johnnie”, NULL) == NULL)
... error ...
mother = john->mother;
father = john->father;
...
soap end recv(&soap);
soap free temp(&soap); // Clean up temporary data but keep deserialized data
}
struct Namespace namespaces[] =
{
{”SOAP-ENV”, ”http://schemas.xmlsoap.org/soap/envelope/”},
{”SOAP-ENC”,”http://schemas.xmlsoap.org/soap/encoding/”},
{”xsi”, ”http://www.w3.org/2001/XMLSchema-instance”},
{”xsd”, ”http://www.w3.org/2001/XMLSchema”},
{”ns”, ”urn:person”}, // Namespace URI of the “Person” data type
{NULL, NULL}
};
It is REQUIRED to either pass NULL to the soap get routine, or a valid pointer to a data structure
that can hold the decoded content. The following example explicitly passes NULL:
john = soap get ns Person(&soap, NULL, ”johnnie”, NULL);
Note: the second NULL parameter indicates that the schema type attribute of the receiving message
can be ignored. The deserializer stores the SOAP contents on the heap, and returns the address.
The allocated storage is released with the soap end call, which removes all temporary and deserialized
data from the heap, or with the soap free temp call, which removes all temporary data only.
Alternatively, the XML content can be decoded within an existing allocated data structure. The
following program fragment decodes the SOAP content in a struct ns Person allocated on the stack:
#include ”soapH.h”
main()
68
{
struct soap soap;
ns Person *mother, *father, john;
soap init(&soap);
soap imode(&soap, SOAP ENC ZLIB); // optional
soap begin(&soap);
soap begin recv(&soap);
soap default ns Person(&soap, &john);
if (soap get ns Person(&soap, &john, ”johnnie”, NULL) == NULL)
... error ...
...
}
struct Namespace namespaces[] =
...
Note the use of soap default ns Person. This routine is generated by the gSOAP stub and skeleton
compiler and assigns default values to the fields of john.
7.5.4
Serializing and Deserializing Class Instances to Streams
C++ applications can define appropriate stream operations on objects for (de)serialization of objects on streams. This is best illustrated with an example. Consider the class
class ns person
{ public:
char *name;
struct soap *soap; // we need this, see below
ns person();
˜ns person();
};
The struct soap member is used to bind the instances to a gSOAP environment for (de)serialization.
We use the gSOAP compiler from the command prompt to generate the class (de)serializers (assuming that person.h contains the class declaration):
soapcpp2 person.h
gSOAP generates the (de)serializers and an instantiation function for the class soap new ns person(struct
soap *soap, int array) to instantiate one or more objects and associate them with a gSOAP environment. The array parameter should be -1 to instantiate one object or should be the number of objects
to instantiate as an array of objects.
#include ”soapH.h”
#include ”ns.nsmap”
...
struct soap *soap = soap new();
ns person *p = soap new ns person(soap, -1);
...
cout << p; // serialize p in XML
69
...
in >> p; // parse XML and deserialize p
...
soap destroy(soap); // deletes p too
soap end(soap);
soap done(soap);
The stream operations are implemented as follows
ostream &operator<<(ostream &o, const ns person &p)
{
if (!p.soap)
return o; // need a gSOAP environment to serialize
p.soap->os = &o;
soap omode(p.soap, SOAP XML GRAPH); // XML tree or graph
p.soap serialize(p.soap);
soap begin send(p.soap);
if (p.soap put(p.soap, ”person”, NULL)
|| soap end send(p.soap))
; // handle I/O error
return o;
}
istream &operator>>(istream &i, ns person &p)
{
if (!p.soap)
return o; // need a gSOAP environment to parse XML and deserialize
p.soap->is = &i;
if (soap begin recv(p.soap)
|| p.soap in(p.soap, NULL, NULL)
|| soap end recv(p.soap))
; // handle I/O error
return i;
}
7.5.5
How to Specify Default Values for Omitted Data
The gSOAP compiler generates soap default functions for all data types. The default values of the
primitive types can be easily changed by defining any of the following macros in the stdsoap2.h file:
#define
#define
#define
#define
#define
#define
#define
#define
#define
#define
#define
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
DEFAULT
DEFAULT
DEFAULT
DEFAULT
DEFAULT
DEFAULT
DEFAULT
DEFAULT
DEFAULT
DEFAULT
DEFAULT
bool
byte
double
float
int
long
LONG64
short
string
time
unsignedByte
70
#define
#define
#define
#define
#define
SOAP
SOAP
SOAP
SOAP
SOAP
DEFAULT
DEFAULT
DEFAULT
DEFAULT
DEFAULT
unsignedInt
unsignedLong
unsignedLONG64
unsignedShort
wstring
Instead of adding these to stdsoap2.h, you can also compile with option -DWITH SOAPDEFS H and
include your definitions in file userdefs.h. The absence of a data value in a receiving SOAP message
will result in the assignment of a default value to a primitive type upon deserialization.
Default values can also be assigned to individual struct and class fields of primitive type. For
example,
struct MyRecord
{
char *name = ”Unknown”;
int value = 9999;
enum Status { active, passive } status = passive;
}
Default values are assigned to the fields on receiving a SOAP/XML message in which the data
values are absent.
Because method requests and responses are essentially structs, default values can also be assigned to
method parameters. The default parameter values do not control the parameterization of C/C++
function calls, i.e. all actual parameters must be present when calling a function. The default
parameter values are used in case an inbound request or response message lacks the XML elements with parameter values. For example, a Web service can use default values to fill-in absent
parameters in a SOAP/XML request:
int ns login(char *uid = ”anonymous”, char *pwd = ”guest”, bool granted);
When the request message lacks uid and pwd parameters, e.g.:
<?xml version="1.0" encoding="UTF-8"?>
<SOAP-ENV:Envelope
xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:SOAP-ENC="http://schemas.xmlsoap.org/soap/encoding/"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:xsd="http://www.w3.org/2001/XMLSchema"
xmlns:ns="http://tempuri.org">
<SOAP-ENV:Body encodingStyle="http://schemas.xmlsoap.org/soap/encoding/">
<ns:login>
</ns:login>
</SOAP-ENV:Body>
</SOAP-ENV:Envelope>
then the service uses the default values. In addition, the default values will show up in the
SOAP/XML request and response message examples generated by the gSOAP compiler.
71
8
Using the gSOAP Stub and Skeleton Compiler
The gSOAP stub and skeleton compiler is invoked from the command line and optionally takes the
name of a header file as an argument or, when the file name is absent, parses the standard input:
soapcpp2 [aheaderfile.h]
where aheaderfile.h is a standard C++ header file. The compiler acts as a preprocessor and produces
C++ source files that can be used to build SOAP client and Web service applications in C++.
The files generated by the compiler are:
File Name
soapStub.h
soapH.h
soapC.cpp
soapClient.cpp
soapServer.cpp
soapClientLib.cpp
soapServerLib.cpp
soapXYZProxy.h
soapXYZProxy.h
soapXYZProxy.cpp
soapXYZObject.h
soapXYZService.h
soapXYZService.cpp
.xsd
.wsdl
.xml
.nsmap
Description
A modified and annotated header file produced from the input header file
Main header file to be included by all client and service sources
Serializers and deserializers for the specified data structures
Client stub routines for remote operations
Service skeleton routines
Client stubs combined with local static (de)serializers
Service skeletons combined with local static (de)serializers
A C++ proxy object (link with soapC.cpp soapClient.cpp)
With option -i: proxy object (link with soapC.cpp and soapXYZProxy.cpp)
With option -i: proxy code
A C++ server object (link with soapC.cpp and soapServer.cpp)
With option -i: server object (link with soapC.cpp and soapXYZService.cpp)
With option -i: server code
An ns.xsd file is generated with an XML Schema for each namespace prefix ns used
by a data structure in the header file input to the compiler, see Section 7.2.8
A ns.wsdl file is generated with an WSDL description for each namespace prefix ns
used by a remote method in the header file input to the compiler, see Section 7.2.8
Several SOAP/XML request and response files are generated. These are example message files are valid provided that sufficient schema namespace directives
are added to the header file or the generated .nsmap namespace table for the
client/service is not modified by hand
A ns.nsmap file is generated for each namespace prefix ns used by a remote method
in the header file input to the compiler, see Section 7.2.8. The file contains a
namespace mapping table that can be used in the client/service sources
Both client and service applications are developed from a header file that specifies the remote
methods. If client and service applications are developed with the same header file, the applications
are guaranteed to be compatible because the stub and skeleton routines use the same serializers and
deserializers to encode and decode the parameters. Note that when client and service applications
are developed together, an application developer does not need to know the details of the internal
SOAP encoding used by the client and service.
The soapClientLib.cpp and soapServerLib.cpp can be used to build (dynamic) client and server libraries.
The serialization routines are local (static) to avoid link symbol conflicts. You must create a separate
library for SOAP Header and Fault handling, as described in Section 18.34.
The following files are part of the gSOAP package and are required to build client and service
applications:
72
File Name
stdsoap2.h
stdsoap2.c
stdsoap2.cpp
8.1
Description
Header file of stdsoap2.cpp runtime library
Runtime C library with XML parser and run-time support routines
Runtime C++ library identical to stdsoap2.c
Compiler Options
The compiler supports the following options:
Option
-1
-2
-C
-S
-L
-a
-c
-d <path>
-e
-h
-i
-I <path>
-l
-m
-n
-p <name>
-s
-t
-v
-w
-x
Description
Use SOAP 1.1 namespaces and encodings (default)
Use SOAP 1.2 namespaces and encodings
Generate client-side code only
Generate server-side code only
Do not generate soapClientLib/soapServerLib
use value of SOAPAction HTTP header to dispatch method at server side
Generate pure C code
Save sources in directory specified by <path>
Generate SOAP RPC encoding style bindings
Print a brief usage message
Generate service proxies and objects inherited from soap struct
Use <path> (or paths separated with ‘:’) for #import
Generate linkable modules (experimental)
Generate Matlabtm code for MEX compiler
When used with -p, enables multi-client and multi-server builds:
Sets compiler option WITH NONAMESPACES, see Section 8.11
Saves the namespace mapping table with name <name> namespaces instead of namespaces
Renames soap serve() into <name> serve() and soap destroy() into <name> destroy()
Save sources with file name prefix <name> instead of “soap”
Generates deserialization code with strict XML validation checks
Generates code to send typed messages (with the xsi:type attribute)
Display version info
Do not generate WSDL and schema files
Do not generate sample XML message files
For example
soapcpp2 -cd ’../projects’ -pmy file.h
Saves the sources:
../projects/myH.h
../projects/myC.c
../projects/myClient.c
../projects/myServer.c
../projects/myStub.h
MS Windows users can use the usual “/” for options, for example:
soapcpp2 /cd ’..\projects’ /pmy file.h
73
Compiler options c, i, n, l, w can be set in the gSOAP header file using the //gsoapopt directive. For
example,
// Generate pure C and do not produce WSDL output:
//gsoapopt cw
int ns myMethod(char*,char**); // takes a string and returns a string
8.2
SOAP 1.1 Versus SOAP 1.2
gSOAP supports SOAP 1.1 by default. SOAP 1.2 support is automatically turned on when the
appropriate SOAP 1.2 namespace is used in the namespace mapping table:
struct Namespace namespaces[] =
{
{”SOAP-ENV”, ”http://www.w3.org/2002/06/soap-envelope”},
{”SOAP-ENC”, ”http://www.w3.org/2002/06/soap-encoding”},
{”xsi”, ”http://www.w3.org/2001/XMLSchema-instance”, ”http://www.w3.org/*/XMLSchemainstance”},
{”xsd”, ”http://www.w3.org/2001/XMLSchema”, ”http://www.w3.org/*/XMLSchema”}, ...
}
gSOAP Web services and clients can automatically switch from SOAP 1.1 to SOAP 1.2 by providing
the SOAP 1.2 namespace as a pattern in the third column of a namespace table:
struct Namespace namespaces[] =
{
{”SOAP-ENV”, ”http://schemas.xmlsoap.org/soap/envelope/”, ”http://www.w3.org/2002/06/soapencoding”},
{”SOAP-ENC”, ”http://schemas.xmlsoap.org/soap/encoding/”, ”http://www.w3.org/2002/06/soapenvelope”},
{”xsi”, ”http://www.w3.org/2001/XMLSchema-instance”, ”http://www.w3.org/*/XMLSchemainstance”},
{”xsd”, ”http://www.w3.org/2001/XMLSchema”, ”http://www.w3.org/*/XMLSchema”}, ...
}
This way, gSOAP Web services can respond to either SOAP 1.1 or SOAP 1.2 requests. gSOAP
will automatically return SOAP 1.2 responses for SOAP 1.2 requests.
The gSOAP soapcpp2 compiler generates a .nsmap file with SOAP-ENV and SOAP-ENC namespace
patterns similar to the above. Since clients issue a send first, they will always use SOAP 1.1
for requests when the namespace table is similar as shown above. Clients can accept SOAP 1.2
responses by inspecting the response message. To restrict gSOAP services and clients to SOAP 1.2
and to generate SOAP 1.2 service WSDLs, use soapcpp2 compiler option -2 to generate SOAP 1.2
conformant .nsmap and .wsdl files.
Caution: SOAP 1.2 does not support partially transmitted arrays. So the
array is meaningless.
offset field of a dynamic
Caution: SOAP 1.2 requires the use of SOAP ENV Code, SOAP ENV Reason, and SOAP ENV Detail
fields in a SOAP ENV Fault fault struct, while SOAP 1.1 uses faultcode, faultstring, and detail fields.
74
Use soap receiver fault subcode(struct soap *soap, const char *subcode, const char *faultstring, const char
*detail) to set a SOAP 1.1/1.2 fault at the server-side with a fault subcode (SOAP 1.2). Use
soap sender fault subcode(struct soap *soap, const char *subcode, const char *faultstring, const char *detail)
to set a SOAP 1.1/1.2 unrecoverable Bad Request fault at the server-side with a fault subcode
(SOAP 1.2).
8.3
The soapdefs.h Header File
The soapdefs.h header file is included in stdsoap2.h when compiling with option -DWITH SOAPDEFS H:
g++ -DWITH SOAPDEFS H -c stdsoap2.cpp
The soapdefs.h file allows users to include definitions and add includes without requiring changes to
stdsoap2.h. For example,
// Contents of soapdefs.h
#include <ostream>
#define SOAP BUFLEN 20480 // use large send/recv buffer
The following header file can now refer to ostream:
extern class ostream; // ostream can’t be (de)serialized, but need to be declared to make it visible
to gSOAP
class ns myClass
{ ...
virtual void print(ostream &s) const; // need ostream here
...
};
See also Section 18.3.
8.4
How to Build Modules and Libraries with the gSOAP #module Directive
The #module directive is used to build modules. A library can be build from a module and linked
with multiple Web services applications. The directive should appear at the top of the header file
and has the following formats:
#module ”name”
and
#module ”name” ”fullname”
75
where name must be a unique short name for the module. The name is case insensitive and MUST
not exceed 4 characters in length. The fullname, when present, represents the full name of the
module.
The rest of the content of the header file includes type declarations and optionally the declarations
of remote methods and SOAP Headers/Faults. When the gSOAP compiler processes the header
file module, it will generate the source codes for a library. The Web services application that uses
the library should use a header file that imports the module with the #import directive.
For example:
/* Contents of module.h */
#module ”test”
long;
char*;
struct ns S
{ ... }
The module.h header file declares a long, char*, and a struct ns X. The module name is ”test”, so
the gSOAP compiler produces a testC.cpp file with the (de)serializers for these types. The testC.cpp
library can be separately compiled and linked with an application that is built from a header file
that imports ”module.h” using #import ”module.h”. You should also compile testClient.cpp when you
want to build a library that includes the remote methods that you defined in the module header
file.
A module MUST be imported into another header file to use it and you cannot use a module alone
to build a SOAP or XML application. That is, the top most header file in the import tree SHOULD
NOT be a module.
When multiple modules are linked, the types that they declare MUST be declared in one module
only to avoid name clashes and link errors. You cannot create two modules that share the same
type declaration and link the modules. When necessary, you should consider creating a module
hierarchy such that types are declared only once and by only one module when these modules must
be linked.
8.5
How to use the gSOAP #import Directive
The #import directive is used to include gSOAP header files into other gSOAP header files for
processing with the gSOAP compiler soapcpp2. The C #include directive cannot be used to include
gSOAP header files. The #include directive is reserved to control the post-gSOAP compilation
process, see 8.6.
The #import directive is used for two purposes: you can use it to include the contents of a header
file into another header file and you can use it to import a module, see 8.4.
An example of the #import directive:
#import ”mydefs.gsoap”
int ns mymethod(xsd string in, xsd int *out);
where ”mydefs.gsoap” is a gSOAP header file that defines xsd string and xsd int:
76
typedef char *xsd string;
typedef int xsd int;
8.6
How to Use #include and #define Directives
The #include and #define directives are normally ignored by the gSOAP compiler. The use of the
directives is enabled with the -i option of the gSOAP compiler, see Section 8.1. However, the gSOAP
compiler will not actually parse the contents of the header files provided by the #include directives
in a header file. Instead, the #include and #define directives will be added to the generated soapH.h
header file before any other header file is included. Therefore, #include and #define directives can
be used to control the C/C++ compilation process of the sources of an application.
The following example header file refers to ostream by including <ostream>:
#include <ostream>
#define WITH COOKIES // use HTTP cookie support (you must compile stdsoap2.cpp with DWITH COOKIES)
#define WITH OPENSSL // enable HTTPS (SSL) support (you must compile stdsoap2.cpp with
-DWITH OPENSSL)
#define SOAP DEFAULT float FLT NAN // use NaN instead of 0.0
extern class ostream; // ostream can’t be (de)serialized, but need to be declared to make it visible
to gSOAP
class ns myClass
{ ...
virtual void print(ostream &s) const; // need ostream here
...
};
This example also uses #define directives for various settings.
Caution: Note that the use of #define in the header file does not automatically result in compiling
stdsoap2.cpp with these directives. You MUST use the -DWITH COOKIES and -DWITH OPENSSL
options when compiling stdsoap2.cpp before linking the object file with your codes. As an alternative,
you can use #define WITH SOAPDEFS H and put the #define directives in the soapdefs.h file.
8.7
Compiling a gSOAP Client
After invoking the gSOAP stub and skeleton compiler on a header file description of a service, the
client application can be compiled on a Linux machine as follows:
g++ -o myclient myclient.cpp stdsoap2.cpp soapC.cpp soapClient.cpp
Or on a Unix machine:
g++ -o myclient myclient.cpp stdsoap2.cpp soapC.cpp soapClient.cpp -lsocket -lxnet -lnsl
(Depending on your system configuration, the libraries libsocket.a, libxnet.a, libnsl.a or dynamic *.so
versions of those libraries are required.)
77
The myclient.cpp file must include soapH.h and must define a global namespace mapping table. A
typical client program layout with namespace mapping table is shown below:
// Contents of file ”myclient.cpp”
#include ”soapH.h”;
...
// A remote method invocation:
soap call some remote method(...);
...
struct Namespace namespaces[] =
{ // {”ns-prefix”, ”ns-name”}
{”SOAP-ENV”, ”http://schemas.xmlsoap.org/soap/envelope/”},
{”SOAP-ENC”, ”http://schemas.xmlsoap.org/soap/encoding/”},
{”xsi”, ”http://www.w3.org/2001/XMLSchema-instance”},
{”xsd”, ”http://www.w3.org/2001/XMLSchema”},
{”ns1”, ”urn:my-remote-method”},
{NULL, NULL}
};
...
A mapping table is generated by the gSOAP compiler that can be used in the source, see Section 7.2.8.
8.8
Compiling a gSOAP Web Service
After invoking the gSOAP stub and skeleton compiler on a header file description of the service,
the server application can be compiled on a Linux machine as follows:
g++ -o myserver myserver.cpp stdsoap2.cpp soapC.cpp soapServer.cpp
Or on a Unix machine:
g++ -o myserver myserver.cpp stdsoap2.cpp soapC.cpp soapServer.cpp -lsocket -lxnet -lnsl
(Depending on your system configuration, the libraries libsocket.a, libxnet.a, libnsl.a or dynamic *.so
versions of those libraries are required.)
The myserver.cpp file must include soapH.h and must define a global namespace mapping table. A
typical service program layout with namespace mapping table is shown below:
// Contents of file ”myserver.cpp”
#include ”soapH.h”;
int main()
{
soap serve(soap new());
}
...
// Implementations of the remote methods as C++ functions
...
78
struct Namespace namespaces[] =
{ // {”ns-prefix”, ”ns-name”}
{”SOAP-ENV”, ”http://schemas.xmlsoap.org/soap/envelope/”},
{”SOAP-ENC”, ”http://schemas.xmlsoap.org/soap/encoding/”},
{”xsi”, ”http://www.w3.org/2001/XMLSchema-instance”},
{”xsd”, ”http://www.w3.org/2001/XMLSchema”},
{”ns1”, ”urn:my-remote-method”},
{NULL, NULL}
};
...
When the gSOAP service is compiled and installed as a CGI application, the soap serve function acts
as a service dispatcher. It listens to standard input and invokes the method via a skeleton routine
to serve a SOAP client request. After the request is served, the response is encoded in SOAP and
send to standard output. The method must be implemented in the server application and the type
signature of the method must be identical to the remote method specified in the header file. That
is, the function prototype in the header file must be a valid prototype of the method implemented
as a C/C++ function.
8.9
Using gSOAP for Creating Web Services and Clients in Pure C
The gSOAP compiler can be used to create pure C Web services and clients. The gSOAP stub and
skeleton compiler soapcpp2 generates .cpp files by default. The compiler generates .c files with the
-c option. However, these files only use C syntax and data types if the header file input to soapcpp2
uses C syntax and data types. For example:
soapcpp2 -c quote.h
gcc -o quote quote.c stdsoap2.c soapC.c soapClient.c
Warnings will be issued by the compiler when C++ class declarations occur in the header file.
8.10
Limitations of gSOAP
gSOAP is SOAP 1.1 and SOAP 1.2 compliant and supports SOAP RPC and document/literal
operations.
From the perspective of the C/C++ language, a few C++ language features are not supported by
gSOAP and these features cannot be used in the specification of SOAP remote methods.
There are certain limitations for the following C++ language constructs:
STL and STL templates The gSOAP compiler supports STL strings std::string and std::wstring
(see Section 10.3.6) and the STL containers std::deque, std::list, std::vector, and std::set, (see
Section 10.11.8).
Templates The gSOAP compiler is a preprocessor that cannot determine the template instantations used by the main program, nor can it generate templated code. You can however
implement containers similar to the STL containers.
79
Multiple inheritance Single class inheritance is supported. Multiple inheritance cannot be supported due to limitations of the SOAP protocol.
Abstract methods A class must be instantiatable to allow decoding of instances of the class.
Directives Directives and pragmas such as #include and #define are interpreted by the gSOAP
compiler. However, the interpretation is different compared to the usual handling of directives,
see Section 8.6. If necessary, a traditional C++ preprocessor can be used for the interpretation
of directives. For example, Unix and Linux users can use “cpp -B” to expand the header file,
e.g. cpp -B myfile.h | soapcpp2. Use the gSOAP #import directive to import gSOAP header
files, see 8.5.
C and C++ programming statements All class methods of a class should be declared within
the class declaration in the header file, but the methods should not be implemented in code.
All class method implementations must be defined within another C++ source file and linked
to the application.
The following data types require some attention to ensure they are serialized:
union types A union data type can not be serialized unless run-time information is associated with
a union in a struct/class as discussed in Section 10.7. An alternative is to use a struct with
a pointer type for each field. Because NULL pointers are not encoded, the resulting encoding
will appear as a union type if only one pointer field is valid (i.e. non-NULL) at the time that
the data type is encoded.
void and void* types The void data type cannot be serialized unless run-time type information is
associated with the pointer using a int type field in the struct/class that contains the void*.
The void* data type is typically used to point to some object or to some array of some type
of objects at run-time. The compiler cannot determine the type of data pointed to and the
size of the array pointed to. A struct or class with a void* field can be augmented to support
the (de)serialization of the void* using a int type field as described in Section 10.9.
Pointers to sequences of elements in memory Any pointer, except for C strings which are
pointers to a sequence of characters, are treated by the compiler as if the pointer points
to only one element in memory at run-time. Consequently, the encoding and decoding
routines will ignore any subsequent elements that follow the first in memory. For the same
reason, arrays of undetermined length, e.g. float a[] cannot be used. gSOAP supports dynamic
arrays using a special type convention, see Section 10.11.
Uninitialized pointers Obviously, all pointers that are part of a data structure must be valid or
NULL to enable serialization of the data structure at run time.
There are a number of programming solutions that can be adopted to circumvent these limitations.
Instead of using void*, a program can in some cases be modified to use a pointer to a known type.
If the pointer is intended to point to different types of objects, a generic base class can be declared
and the pointer is declared to point to the base class. All the other types are declared to be derived
classes of this base class. For pointers that point to a sequence of elements in memory dynamic
arrays should be used instead, see 10.11.
80
8.11
Compile Time Flags
The following macros (#defines) can be used to enable certain optional features:
Macro
WITH SOAPDEFS H
WITH COOKIES
WITH OPENSSL
WITH IPV6
WITH TCPFIN
WITH FASTCGI
WITH GZIP
WITH ZLIB
WITH UDP
WITH FAST
WITH NOIO
WITH NOHTTP
WITH LEAN
WITH LEANER
WITH NONAMESPACES
WITH
WITH
WITH
WITH
NOEMPTYSTRUCT
NOGLOBAL
CDATA
CASEINSENSITIVETAGS
Description
includes the soapdefs.h file for custom settings, see Section 8.3
enables HTTP cookies, see Sections 18.26 18.27
enables OpenSSL, see Sections 18.20 18.19
enables IPv6 support
use TCP FIN after sends when socket is ready to close
enables FastCGI, see Sections 18.29
enables gzip and deflate compression, see Section 18.25
enables deflate compression only, see Section 18.25
enables UDP support (SOAP-over-UDP), see Section 17
(obsolete)
eliminates need for file IO and BSD socket library, see Section 18.31
eliminates HTTP stack to reduce code size
creates a small-footprint executable, see Section 18.30
creates an even smaller footprint executable, see Section 18.30
omit initialization of soap struct with global namespaces table
and you MUST explicitly initialize soap.namespaces to point to a table
see also Section 9.4
inserts a dummy member in empty structs to enable compilation
omit SOAP Header and Fault serialization code
retain the parsed CDATA sections in literal XML strings (no conversion)
enable case insensitive XML parsing
Caution: it is important that the macros MUST be consistently defined to compile the sources,
such as stdsoap2.cpp, soapC.cpp, soapClient.cpp, soapServer.cpp, and all application sources that include
stdsoap2.h or soapH.h. If the macros are not consistently used, the application will crash due to a
mismatches in the declaration and access of the gSOAP environment.
8.12
Run Time Flags
gSOAP provides flags to control the input and output mode settings at runtime. These flags are
divided into four categories: transport (IO), content encoding (ENC), XML marshalling (XML),
and C/C++ data mapping (C).
Although gSOAP is fully SOAP 1.1 compliant, some SOAP implementations may have trouble
accepting multi-reference data and/or require explicit nil data so these flags can be used to put
gSOAP in “safe mode”. In addition, the embedding (or inlining) of multi-reference data is adopted
in the SOAP 1.2 specification, which gSOAP automatically supports when handling with SOAP
1.2 messages. The flags are:
81
Flag
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
SOAP
IO FLUSH
IO BUFFER
IO STORE
IO CHUNK
IO LENGTH
IO KEEPALIVE
IO UDP
ENC XML
ENC DIME
ENC MIME
ENC MTOM
ENC ZLIB
ENC SSL
XML INDENT
XML CANONICAL
XML STRICT
XML TREE
XML GRAPH
XML NIL
C NOIOB
C UTFSTRING
C MBSTRING
C NILSTRING
Description
Disable buffering and flush output (default for all file-based output)
Enable buffering (default for all socket-oriented connections)
Store entire message to calculate HTTP content length
Use HTTP chunking
Internal flag: require apriori calculation of content length
Attempt to keep socket connections alive (open)
Use UDP (datagram) transport, maximum message length is SOAP BUFLEN
Use plain XML encoding without HTTP headers (useful with SOAP ENC ZLIB)
Use DIME encoding (automatic when DIME attachments are used)
Use MIME encoding (activate using soap set mime)
Use MTOM XOP attachments (instead of DIME)
Compress encoding with Zlib (deflate or gzip format)
Encrypt encoding with SSL (automatic with ”https:” endpoints)
Produce indented XML output
Produce canonical XML output
XML strict validation
Serialize data as XML trees (no multi-ref, duplicate data when necessary)
Serialize data as an XML graph with inline multi-ref (SOAP 1.2 default)
Serialize NULL data as xsi:nil elements (omit by default)
Do not fault with SOAP IOB
(De)serialize 8-bit strings “as is” (strings MUST have UTF-8 encoded content)
Activate multibyte character support (depends on locale)
Serialize empty strings as nil (ommited element)
The flags can be selectively turned on/off at any time, for example when multiple Web services are
accessed by a client that require special treatment.
All flags are orthogonal, except SOAP IO FLUSH, SOAP IO BUFFER, SOAP IO STORE, and SOAP IO CHUNK
which are enumerations and only one of these I/O flags can be used. Also the XML serialization
flags SOAP XML TREE and SOAP XML GRAPH should not be mixed.
The flags control the inbound and outbound message transport, encoding, and (de)serialization.
The following functions are used to set and reset the flags for input and output modes:
Function
soap init2(struct soap *soap, int imode, int omode)
soap imode(struct soap *soap, int imode)
soap omode(struct soap *soap, int omode)
soap set imode(struct soap *soap, int imode)
soap set omode(struct soap *soap, int omode)
soap clr imode(struct soap *soap, int omode)
soap clr omode(struct soap *soap, int omode)
Description
Initialize the runtime and set flags
Set all input mode flags
Set all output mode flags
Enable input mode flags
Enable output mode flags
Disable input mode flags
Disable output mode flags
The default setting is SOAP IO DEFAULT for both input and output modes.
For example
struct soap soap;
soap init2(&soap, SOAP IO KEEPALIVE,
SOAP IO KEEPALIVE|SOAP ENC ZLIB|SOAP XML TREE|SOAP XML CANONICAL);
if (soap call ns myMethod(&soap, ...))
...
82
sends a compressed client request with keep-alive enabled and all data serialized as canonical XML
trees.
In many cases, setting the input mode will have no effect, especially with HTTP transport because gSOAP will determine the optimal input buffering and the encoding used for an inbound
message. The flags that have an effect on handling inbound messages are SOAP IO KEEPALIVE,
SOAP ENC SSL (but automatic when ”https:” endpoints are used or soap ssl accept), SOAP C NOIOB,
SOAP C UTFSTRING, and SOAP C MBSTRING.
Caution: The SOAP XML TREE serialization flag can be used to improve interoperability with
SOAP implementations that are not fully SOAP 1.1 compliant. However, a tree serialization will
duplicate data when necessary and will crash the serializer for cyclic data structures.
8.13
Memory Management
Understanding gSOAP’s run-time memory management is important to optimize client and service
applications by eliminating memory leaks and/or dangling references.
There are two forms of dynamic (heap) allocations made by gSOAP’s runtime for serialization and
deserialization of data. Temporary data is created by the runtime such as hash tables to keep
pointer reference information for serialization and hash tables to keep XML id/href information for
multi-reference object deserialization. Deserialized data is created upon receiving SOAP messages.
This data is stored on the heap and requires several calls to the malloc library function to allocate
space for the data and new to create class instances. All such allocations are tracked by gSOAP’s
runtime by linked lists for later deallocation. The linked list for malloc allocations uses some extra
space in each malloced block to form a chain of pointers through the malloced blocks. A separate
malloced linked list is used to keep track of class instance allocations.
gSOAP does not enforce a deallocation policy and the user can adopt a deallocation policy that
works best for a particular application. As a consequence, deserialized data is never deallocated by
the gSOAP runtime unless the user explicitly forces deallocation by calling functions to deallocate
data collectively or individually.
The deallocation functions are:
Function Call
soap destroy(struct soap *soap)
soap end(struct soap *soap)
soap free temp(struct soap *soap)
soap dealloc(struct soap *soap, void *p)
soap delete(struct soap *soap, void *p)
soap unlink(struct soap *soap, void *p)
soap done(struct soap *soap)
soap free(struct soap *soap)
Description
Remove all dynamically allocated C++ objects.
must be called before soap end()
Remove temporary data and deserialized data except
class instances
Instead of soap destroy and soap end:
remove temporary data only
Remove malloced data at p. When p==NULL: remove all
dynamically allocated (deserialized) data except class instances
Remove class instance at p. When p==NULL: remove all
dynamically allocated (deserialized) class instances
(this is identical to calling soap destroy(struct soap *soap))
Unlink data/object at p from gSOAP’s deallocation chain
so gSOAP won’t deallocate it
Detach context (reset runtime environment)
Detach and free context (allocated with soap new)
83
Temporary data (i.e. the hash tables) are automatically removed with calls to the soap free temp
function which is made within soap end and soap done or when the next call to a stub or skeleton
routine is made to send a message or receive a message. Deallocation of non-class based data is
straightforward: soap end removes all dynamically allocated deserialized data (data allocated with
soap malloc. That is, when the client/service application does not use any class instances that are
(de)marshalled, but uses structs, arrays, etc., then calling the soap end function is safe to remove
all deserialized data. The function can be called after processing the deserialized data of a remote
method call or after a number of remote method calls have been made. The function is also typically
called after soap serve, when the service finished sending the response to a client and the deserialized
client request data can be removed.
Individual data objects can be unlinked from the deallocation chain if necessary, to prevent deallocation by the collective soap end or soap destroy functions.
8.13.1
Memory Management Policies
There are three situations to consider for memory deallocation policies for class instances:
1. the program code deletes the class instances and the class destructors in turn SHOULD delete
and free any dynamically allocated data (deep deallocation) without calling the soap end and
soap destroy functions,
2. or the class destructors SHOULD NOT deallocate any data and the soap end and soap destroy
functions can be called to remove the data.
3. or the class destructors SHOULD mark their own deallocation and mark the deallocation
of any other data deallocated by it’s destructors by calling the soap unlink function. This
allows soap destroy and soap end to remove the remaining instances and data without causing
duplicate deallocations.
It is advised to use pointers to class instances that are used within other structs and classes to avoid
the creation of temporary class instances during deserialization. The problem with temporary class
instances is that the destructor of the temporary may affect data used by other instances through
the sharing of data parts accessed with pointers. Temporaries and even whole copies of class
instances can be created when deserializing SOAP multi-referenced objects. A dynamic array of
class instances is similar: temporaries may be created to fill the array upon deserialization. To
avoid problems, use dynamic arrays of pointers to class instances. This also enables the exchange
of polymorphic arrays when the elements are instances of classes in an inheritance hierarchy. In
addition, allocate data and class instances with soap malloc and soap new X functions (more details
below).
To summarize, it is advised to pass class data types by pointer to a remote method. For example:
class X { ... };
ns remoteMethod(X *in, ...);
Response elements that are class data types can be passed by reference, as in:
84
class X { ... };
class ns remoteMethodResponse { ... };
ns remoteMethod(X *in, ns remoteMethodResponse &out);
But dynamic arrays declared as class data types should use a pointer to a valid object that will be
overwritten when the function is called, as in:
typedef int xsd int;
class X { ... };
class ArrayOfint { xsd int * ptr; int size; };
ns remoteMethod(X *in, ArrayOfint *out);
Or a reference to a valid or NULL pointer, as in:
typedef int xsd int;
class X { ... };
class ArrayOfint { xsd int * ptr; int size; };
ns remoteMethod(X *in, ArrayOfint *&out);
The gSOAP memory allocation functions can be used in client and/or service code to allocate
temporary data that will be automatically deallocated. These functions are:
Function Call
void *soap malloc(struct soap *soap, size t n)
Class *soap new Class(struct soap *soap, int n)
Description
return pointer to n bytes
instantiate n Class objects
The soap new X functions are generated by the gSOAP compiler for every class X in the header file.
Parameter n MUST be -1 to instantiate a single object, or larger or equal to 0 to instantiate an
array of n objects.
Space allocated with soap malloc will be released with the soap end and soap dealloc functions. Objects
instantiated with soap new X(struct soap*) are removed altogether with soap destroy(struct soap*).
Individual objects instantiated with soap new X are removed with soap delete X(struct soap*, X*). For
example, the following service uses temporary data in the remote method implementation:
int main()
{ ...
struct soap soap;
soap init(&soap);
soap serve(&soap);
soap end(&soap);
...
}
An example remote method that allocates a temporary string is:
int ns itoa(struct soap *soap, int i, char **a)
{
*a = (char*)soap malloc(soap, 11);
sprintf(*a, ”%d”, i);
return SOAP OK;
}
85
This temporary allocation can also be used to allocate strings for the SOAP Fault data structure.
For example:
int ns mymethod(...)
{ ...
if (exception)
{
char *msg = (char*)soap malloc(soap, 1024); // allocate temporary space for detailed message
sprintf(msg, ”...”, ...); // produce the detailed message
return soap receiver fault(soap, ”An exception occurred”, msg); // return the server-side fault
}
...
}
Use soap receiver fault(struct soap *soap, const char *faultstring, const char *detail) to set a SOAP 1.1/1.2
fault at the server-side. Use soap sender fault(struct soap *soap, const char *faultstring, const char *detail)
to set a SOAP 1.1/1.2 unrecoverable Bad Request fault at the server-side. Sending clients are not
supposed to retry messages after a Bad Request, while errors at the receiver-side indicate temporary
problems.
The above functions do not include a SOAP 1.2 Subcode element. To include Subcode element, use
soap receiver fault subcode(struct soap *soap, const char *subcode, const char *faultstring, const char *detail)
to set a SOAP 1.1/1.2 fault with Subcode at the server-side. Use soap sender fault subcode(struct
soap *soap, const char *subcode, const char *faultstring, const char *detail) to set a SOAP 1.1/1.2
unrecoverable Bad Request fault with Subcode at the server-side.
gSOAP provides a function to duplicate a string into gSOAP’s memory space:
char *soap strdup(struct soap *soap, const char *s)
The function allocates space for s with soap malloc, copies the string, and returns a pointer to the
duplicated string. When s is NULL, the function does not allocate and copy the string and returns
NULL.
8.13.2
Intra-Class Memory Management
When a class declaration has a struct soap * field, this field will be set to point to the current gSOAP
run-time environment by gSOAP’s deserializers and by the soap new Class functions. This simplifies
memory management for class instances. The struct soap* pointer is implicitly set by the gSOAP
deserializer for the class or explicitly by calling the soap new X function for class X. For example:
class Sample
{ public:
struct soap *soap; // reference to gSOAP’s run-time
...
Sample();
˜Sample();
};
The constructor and destructor for class Sample are:
86
Sample::Sample()
{ this->soap = NULL;
}
Sample::˜Sample()
{ soap unlink(this->soap, this);
}
The soap unlink() call removes the object from gSOAP’s deallocation chain. In that way, soap destroy
can be safely called to remove all class instances. The following code illustrates the explicit creation
of a Sample object and cleanup:
struct soap *soap = soap new(); // new gSOAP runtime
Sample *obj = soap new Sample(soap, -1); // new Sample object with obj->soap set to runtime
...
delete obj; // also calls soap unlink to remove obj from the deallocation chain
soap destroy(soap); // deallocate all (other) class instances
soap end(soap); // clean up
Here is another example:
class ns myClass
{ ...
struct soap *soap; // set by soap new ns myClass()
char *name;
void setName(const char *s);
...
};
Calls to soap new ns myClass(soap, n) will set the soap field in the class instance to the current
gSOAP environment. Because the deserializers invoke the soap new functions, the soap field of the
ns myClass instances are set as well. This mechanism is convenient when Web Service methods
need to return objects that are instantiated in the methods. For example
int ns myMethod(struct soap *soap, ...)
{
ns myClass *p = soap new ns myClass(soap, -1);
p->setName(”SOAP”);
return SOAP OK;
}
void ns myClass::ns setName(const char *s)
{
if (soap)
name = (char*)soap malloc(soap, strlen(s)+1);
else
name = (char*)malloc(strlen(s)+1);
strcpy(name, s);
}
ns myClass::ns myClass()
{
soap = NULL;
87
name = NULL;
}
ns myClass::˜ns myClass()
{
if (!soap && name) free(name);
soap unlink(soap, this);
}
Calling soap destroy right after soap serve in the Web Service will destroy all dynamically allocated
class instances.
8.14
Debugging
To activate message logging for debugging, un-comment the #define DEBUG directive in stdsoap2.h.
Compile the client and/or server applications as described above (or simply use g++ -DDEBUG ...
to compile with debugging activated). When the client and server applications run, they will log
their activity in three separate files:
File
SENT.log
RECV.log
TEST.log
Description
The SOAP content transmitted by the application
The SOAP content received by the application
A log containing various activities performed by the application
Caution: The client and server applications may run slow due to the logging activity.
Caution: When installing a CGI application on the Web with debugging activated, the log files may
sometimes not be created due to file access permission restrictions imposed on CGI applications.
To get around this, create empty log files with universal write permissions. Be careful about the
security implication of this.
You can test a service CGI application without deploying it on the Web. To do this, create a client
application for the service and activate message logging by this client. Remove any old SENT.log file
and run the client (which connects to the Web service or to another dummy, but valid address) and
copy the SENT.log file to another file, e.g. SENT.tst. Then redirect the SENT.tst file to the service
CGI application. For example,
myservice.cgi < SENT.tst
This should display the service response on the terminal.
The file names of the log files and the logging activity can be controlled at the application level.
This allows the creation of separate log files by separate services, clients, and threads. For example,
the following service logs all SOAP messages (but no debug messages) in separate directories:
struct soap soap;
soap init(&soap);
...
soap set recv logfile(&soap, ”logs/recv/service12.log”); // append all messages received in /logs/recv/service12.log
soap set sent logfile(&soap, ”logs/sent/service12.log”); // append all messages sent in /logs/sent/service12.log
88
soap set test logfile(&soap, NULL); // no file name: do not save debug messages
...
soap serve(&soap);
...
Likewise, messages can be logged for individual client-side remote method calls.
8.15
Required Libraries
• The socket library is essential and requires the inclusion of the appropriate libraries with the
compile command for Sun Solaris systems:
g++ -o myclient myclient.cpp stdsoap2.cpp soapC.cpp soapClient.cpp -lsocket -lxnet -lnsl
These library loading options are not required with Linux.
• The gSOAP runtime uses the math library for the NaN, INF, and -INF floating point representations. The library is not strictly necessary and the <math.h> header file import can be
commented out from the stdsoap2.h header file. The application can be linked without the -lm
math library e.g. under Sun Solaris:
g++ -o myclient myclient.cpp stdsoap2.cpp soapC.cpp soapClient.cpp -lsocket -lxnet -lnsl
9
The gSOAP Remote Method Specification Format
A SOAP remote method is specified as a C/C++ function prototype in a header file. The function
is REQUIRED to return int, which is used to represent a SOAP error code, see Section 9.2. Multiple
remote methods MAY be declared together in one header file.
The general form of a SOAP remote method specification is:
[int] [namespace prefix ]method name([inparam1, inparam2, ...,] outparam);
where
namespace prefix
is the optional namespace prefix of the method (see identifier translation rules 9.3)
method name it the remote method name (see identifier translation rules 9.3)
inparam is the declaration of an input parameter of the remote method
outparam is the declaration of the output parameter of the remote method
This simple form can only pass a single, non-struct and non-class type output parameter. See 9.1 for
passing multiple output parameters. The name of the declared function namespace prefix method name
must be unique and cannot match the name of a struct, class, or enum declared in the same header
file.
The method request is encoded in SOAP as an XML element and the namespace prefix, method
name, and input parameters are encoded using the format:
89
<[namespace-prefix:]method name xsi:type="[namespace-prefix:]method name>
<inparam-name1 xsi:type="...">...</inparam-name1>
<inparam-name2 xsi:type="...">...</inparam-name2>
...
</[namespace-prefix:]method name>
where the inparam-name accessors are the element-name representations of the inparam parameter
name declarations, see Section 9.3. (The optional parts are shown enclosed in [].)
The XML response by the Web service is of the form:
<[namespace-prefix:]method-nameResponse xsi:type="[namespace-prefix:]method-nameResponse>
<outparam-name xsi:type="...">...</outparam-name>
</[namespace-prefix:]method-nameResponse>
where the outparam-name accessor is the element-name representation of the outparam parameter
name declaration, see Section 9.3. By convention, the response element name is the method name
ending in Response. See 9.1 on how to change the declaration if the service response element name
is different.
The gSOAP stub and skeleton compiler generates a stub routine for the remote method. This stub
is of the form:
int soap call [namespace prefix ]method name(struct soap *soap, char *URL, char *action, [inparam1,
inparam2, ...,] outparam);
This proxy can be called by a client application to perform the remote method call.
The gSOAP stub and skeleton compiler generates a skeleton routine for the remote method. The
skeleton function is:
int soap serve [namespace prefix ]method name(struct soap *soap);
The skeleton routine, when called by a service application, will attempt to serve a request on
the standard input. If no request is present or if the request does not match the method name,
SOAP NO METHOD is returned. The skeleton routines are automatically called by the generated
soap serve routine that handles all requests.
9.1
Remote Method Parameter Passing
The input parameters of a remote method MUST be passed by value. Input parameters cannot be
passed by reference with the & reference operator, but an input parameter value MAY be passed
by a pointer to the data. Of course, passing a pointer to the data is preferred when the size of the
data of the parameter is large. Also, to pass instances of (derived) classes, pointers to the instance
need to be used to avoid passing the instance by value which requires a temporary and prohibits
passing derived class instances. When two input parameter values are identical, passing them using
a pointer has the advantage that the value will be encoded only once as multi-reference (hence, the
parameters are aliases). When input parameters are passed using a pointer, the data pointed to
will not be modified by the remote method and returned to the caller.
90
The output parameter MUST be passed by reference using & or by using a pointer. Arrays are
passed by reference by default and do not require the use of the reference operator &.
The input and output parameter types have certain limitations, see Section 8.10
If the output parameter is a struct or class type, it is considered a SOAP remote method response
element instead of a simple output parameter value. That is, the name of the struct or class is
the name of the response element and the struct or class fields are the output parameters of the
remote method, see also 7.1.7. Hence, if the output parameter has to be a struct or class, a response
struct or class MUST be declared as well. In addition, if a remote method returns multiple output
parameters, a response struct or class MUST be declared. By convention, the response element is
the remote method name ending with “Response”.
The general form of a response element declaration is:
struct [namespace prefix ]response element name
{
outparam1;
outparam2;
...
};
where
namespace prefix
is the optional namespace prefix of the response element (see identifier translation
rules 9.3)
response element name it the name of the response element (see identifier translation rules 9.3)
outparam is the declaration of an output parameter of the remote method
The general form of a remote method specification with a response element declaration for (multiple)
output parameters is:
[int] [namespace prefix ]method name([inparam1, inparam2, ...,] struct [namespace prefix ]response element name
{outparam1[, outparam2, ...]} &anyparam);
The choice of name for anyparam has no effect on the SOAP encoding and decoding and is only used
as a place holder for the response.
The method request is encoded in SOAP as an independent element and the namespace prefix,
method name, and input parameters are encoded using the format:
<[namespace-prefix:]method-name xsi:type="[namespace-prefix:]method-name>
<inparam-name1 xsi:type="...">...</inparam-name1>
<inparam-name2 xsi:type="...">...</inparam-name2>
...
</[namespace-prefix:]method-name>
where the inparam-name accessors are the element-name representations of the inparam parameter
name declarations, see Section 9.3. (The optional parts resulting from the specification are shown
enclosed in [].)
91
The method response is expected to be of the form:
<[namespace-prefix:]response-element-name xsi:type="[namespace-prefix:]response-element-name>
<outparam-name1 xsi:type="...">...</outparam-name1>
<outparam-name2 xsi:type="...">...</outparam-name2>
...
</[namespace-prefix:]response-element-name>
where the outparam-name accessors are the element-name representations of the outparam parameter
name declarations, see Section 9.3. (The optional parts resulting from the specification are shown
enclosed in [].)
The input and/or output parameters can be made anonymous, which allows the deserialization of
requests/responses with different parameter names as is endorsed by the SOAP 1.1 specification,
see Section 7.1.13.
9.2
Error Codes
The error codes returned by the stub and skeleton routines are listed below.
92
Code
SOAP OK
SOAP CLI FAULT*
SOAP SVR FAULT*
SOAP TAG MISMATCH
SOAP TYPE
SOAP SYNTAX ERROR
SOAP NO TAG
SOAP IOB
SOAP MUSTUNDERSTAND*
SOAP NAMESPACE
SOAP OBJ MISMATCH
SOAP FATAL ERROR
SOAP FAULT
SOAP NO METHOD
SOAP NO DATA
SOAP GET METHOD
SOAP EOM
SOAP NULL
SOAP MULTI ID
SOAP MISSING ID
SOAP HREF
SOAP UDP ERROR
SOAP TCP ERROR
SOAP HTTP ERROR
SOAP SSL ERROR
SOAP ZLIB ERROR
SOAP PLUGIN ERROR
SOAP MIME ERROR
SOAP DIME ERROR
SOAP DIME END
SOAP DIME HREF
SOAP DIME MISMATCH
SOAP VERSIONMISMATCH*
SOAP DATAENCODINGUNKNOWN
SOAP REQUIRED
SOAP PROHIBITED
SOAP OCCURS
SOAP LENGTH
SOAP EOF
Description
No error
The service returned a client fault (SOAP 1.2 Sender fault)
The service returned a server fault (SOAP 1.2 Receiver fault)
An XML element didn’t correspond to anything expected
An XML Schema type mismatch
An XML syntax error occurred on the input
Begin of an element expected, but not found
Array index out of bounds
An element needs to be ignored that need to be understood
Namespace name mismatch (validation error)
Mismatch in the size and/or shape of an object
Internal error
An exception raised by the service
The dispatcher did not find a matching operation for the request
No data in HTTP message
HTTP GET operation not handled, see Section 18.10
Out of memory
An element was null, while it is not supposed to be null
Multiple occurrences of the same element ID on the input
Element ID missing for an HREF on the input
Reference to object is incompatible with the object refered to
Message too large to store in UDP packet
A connection error occured
An HTTP error occured
An SSL error occured
A Zlib error occured
Failed to register plugin
MIME parsing error
DIME parsing error
End of DIME error
DIME attachment has no href from SOAP body
(and no DIME callbacks were defined to save the attachment)
DIME version/transmission error
SOAP version mismatch or no SOAP message
SOAP 1.2 DataEncodingUnknown fault
Attributed required validation error
Attributed prohibited validation error
Element minOccurs/maxOccurs validation error
Element length validation error
Unexpected end of file, no input, or timeout while receiving data
The error codes that are returned by a stub routine (proxy) upon receiving a SOAP Fault from
the server are marked (*). The remaining error codes are generated by the proxy itself as a result
of problems with a SOAP payload. The error code is SOAP OK when the remote method call was
successful (the SOAP OK predefined constant is guaranteed to be 0). The error code is also stored
in soap.error, where soap is a variable that contains the current runtime environment. The function
soap print fault(struct soap *soap, FILE *fd) can be called to display an error message on fd where
current value of the soap.error variable is used by the function to display the error. The function
soap print fault location(struct soap *soap, FILE *fd) prints the location of the error if the error is a
result from parsing XML.
93
A remote method implemented in a SOAP service MUST return an error code as the function’s
return value. SOAP OK denotes success and SOAP FAULT denotes an exception. The exception
details can be assigned with the soap receiver fault(struct soap *soap, const char *faultstring, const
char *detail) which sets the strings soap.fault->faultstring and soap.fault->detail for SOAP 1.1, and
soap.fault->SOAP ENV Reason and soap.fault->SOAP ENV Detail for SOAP 1.2, where soap is a variable that contains the current runtime environment, see Section 11. A receiver error indicates
that the service can’t handle the request, but can possibly recover from the error. To return an
unrecoverable error, use soap receiver fault(struct soap *soap, const char *faultstring, const char *detail).
To return a HTTP error code a service method can simply return the HTTP error code number.
For example, return 404; returns a ”404 Not Found” HTTP error back to the client. The soap.error
is set to the HTTP error code at the client side. The HTTP 1.1 error codes are:
#
201
202
203
204
205
206
300
301
302
303
304
305
307
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
500
501
502
503
504
505
Error
Created
Accepted
Non-Authoritative Information
No Content
Reset Content
Partial Content
Multiple Choices
Moved Permanently
Found
See Other
Not Modified
Use Proxy
Temporary Redirect
Bad Request
Unauthorized
Payment Required
Forbidden
Not Found
Method Not Allowed
Not Acceptable
Proxy Authentication Required
Request Time-out
Conflict
Gone
Length Required
Precondition Failed
Request Entity Too Large
Request-URI Too Large
Unsupported Media Type
Requested range not satisfiable
Expectation Failed
Internal Server Error
Not Implemented
Bad Gateway
Service Unavailable
Gateway Time-out
HTTP Version not supported
94
The error codes are given for informational purposes only. The HTTP protocol requires the proper
actions after an error is issued. gSOAP’s HTTP 1.0/1.1 handling is automatic.
9.3
C/C++ Identifier Name to XML Name Translations
One of the “secrets” behind the power and flexibility of gSOAP’s encoding and decoding of remote
method names, class names, type identifiers, and struct or class fields is the ability to specify
namespace prefixes with these names that are used to denote their encoding style. More specifically,
a C/C++ identifier name of the form
[namespace prefix ]element name
will be encoded in XML as
<[namespace-prefix:]element-name ...>
The underscore pair ( ) separates the namespace prefix from the element name. Each namespace
prefix has a namespace URI specified by a namespace mapping table 9.4, see also Section 7.1.2.
The namespace URI is a unique identification that can be associated with the remote methods and
data types. The namespace URI disambiguates potentially identical remote method names and
data type names used by disparate organizations.
XML element names are NCNames (restricted strings) that MAY contain hyphens, dots, and
underscores. The special characters in the XML element names of remote methods, structs,
classes, typedefs, and fields can be controlled using the following conventions: A single underscore
in a namespace prefix or identifier name is replaced by a hyphen (-) in the XML element name. For
example, the identifier name SOAP ENC ur type is represented in XML as SOAP-ENC:ur-type. The
sequence DOT is replaced by a dot (.), and the sequence USCORE is replaced by an underscore ( )
in the corresponding XML element name. For example:
class n s biz DOTcom
{
char *n s biz USCOREname;
};
is encoded in XML as:
<n-s:biz.com xsi:type="n-s:biz.com">
<n-s:biz name xsi:type="string">Bizybiz</n-s:biz name>
</n-s:biz.com>
Trailing underscores of an identifier name are not translated into the XML representation. This is
useful when an identifier name clashes with a C++ keyword. For example, return is often used
as an accessor name in a SOAP response element. The return element can be specified as return
in the C++ source code. Note that XML should be treated as case sensitive, so the use of e.g.
Return may not always work to avoid a name clash with the return keyword. The use of trailing
underscores also allows for defining structs and classes with essentially the same XML Schema type
name, but that have to be distinguished as seperate C/C++ types.
95
For decoding, the underscores in identifier names act as wildcards. An XML element is parsed and
matches the name of an identifier if the name is identical to the element name (case insensitive)
and the underscores in the identifier name are allowed to match any character in the element
name. For example, the identifier name I want soap fun the bea DOTcom matches the element
name I-want:[email protected].
9.4
Namespace Mapping Table
A namespace mapping table MUST be defined by clients and service applications. The mapping
table is used by the serializers and deserializers of the stub and skeleton routines to produce a valid
SOAP payload and to validate an incoming SOAP payload. A typical mapping table is shown
below:
struct Namespace namespaces[] =
{ // {”ns-prefix”, ”ns-name”}
{”SOAP-ENV”, ”http://schemas.xmlsoap.org/soap/envelope/”}, // MUST be first
{”SOAP-ENC”, ”http://schemas.xmlsoap.org/soap/encoding/”}, // MUST be second
{”xsi”, ”http://www.w3.org/2001/XMLSchema-instance”}, // MUST be third
{”xsd”, ”http://www.w3.org/2001/XMLSchema”}, // Required for XML Schema types
{”ns1”, ”urn:my-service-URI”}, // The namespace URI of the remote methods
{NULL, NULL} // end of table
};
Each namespace prefix used by a identifier name in the header file specification (see Section 9.3)
MUST have a binding to a namespace URI in the mapping table. The end of the namespace mapping table MUST be indicated by the NULL pair. The namespace URI matching is case insensitive.
A namespace prefix is distinguished by the occurrence of a pair of underscores ( ) in an identifier.
An optional namespace pattern MAY be provided with each namespace mapping table entry. The
patterns provide an alternative namespace matching for the validation of decoded SOAP messages.
In this pattern, dashes (-) are single-character wildcards and asterisks (*) are multi-character wildcards. For example, to decode different versions of XML Schema type with different authoring
dates, four dashes can be used in place of the specific dates in the namespace mapping table
pattern:
struct Namespace namespaces[] =
{ // {”ns-prefix”, ”ns-name”, ”ns-name validation pattern”}
...
{”xsi”, ”http://www.w3.org/2001/XMLSchema-instance”, ”http://www.w3.org/----/XMLSchemainstance”},
{”xsd”, ”http://www.w3.org/2001/XMLSchema”, ”http://www.w3.org/----/XMLSchema”},
...
Or alternatively, asterisks can be used as wildcards for multiple characters:
struct Namespace namespaces[] =
{ // {”ns-prefix”, ”ns-name”, ”ns-name validation pattern”}
96
...
{”xsi”, ”http://www.w3.org/2001/XMLSchema-instance”, ”http://www.w3.org/*/XMLSchemainstance”},
{”xsd”, ”http://www.w3.org/2001/XMLSchema”, ”http://www.w3.org/*/XMLSchema”},
...
A namespace mapping table is automatically generated together with a WSDL file for each namespace prefix that is used for a remote method in the header file. This namespace mapping table
has entries for all namespace prefixes. The namespace URIs need to be filled in. These appear as
http://tempuri.org in the table. See Section 18.2 on how to specify the namespace URIs in the header
file.
For decoding elements with namespace prefixes, the namespace URI associated with the namespace
prefix (through the xmlns attribute of an XML element) is searched from the beginning to the end
in a namespace mapping table, and for every row the following tests are performed as part of the
validation process:
1. the string in the second column matches the namespace URI (case insensitive)
2. the string in the optional third column matches the namespace URI (case insensitive), where
- is a one-character wildcard and * is a multi-character wildcard
When a match is found, the namespace prefix in the first column of the table is considered semantically identical to the namespace prefix used by the XML element to be decoded, though the prefix
names may differ. A service will respond with the namespace that it received from a client in case
it matches a pattern in the third column.
For example, let’s say we have the following structs:
struct a elt { ... };
struct b elt { ... };
struct k elt { ... };
and a namespace mapping table in the program:
struct Namespace namespaces[] =
{ // {”ns-prefix”, ”ns-name”, ”ns-name validation pattern”}
...
{”a”, ”some uri”},
{”b”, ”other uri”},
{”c”, ”his uri”, ”* uri”},
...
Then, the following XML elements will match the structs:
<n:elt xmlns:n="some URI">
...
<m:elt xmlns:m="other URI">
...
matches the struct name a elt
matches the struct name b elt
97
<k:elt xmlns:k="my URI">
...
matches the struct name c elt
The response of a service to a client request that uses the namespaces listed above, will include my
URI for the name space of element k.
It is possible to use a number of different namespace tables and select the one that is appropriate.
For example, an application might contact many different Web services all using different namespace
URIs. If all the URIs are stored in one table, each remote method invocation will dump the whole
namespace table in the SOAP payload. There is no technical problem with that, but it can be ugly
when the table is large. To use different namespace tables, declare a pointer to a table and set the
pointer to a particular table before remote method invocation. For example:
struct Namespace namespacesTable1[] = { ... };
struct Namespace namespacesTable2[] = { ... };
struct Namespace namespacesTable3[] = { ... };
struct Namespace *namespaces;
...
struct soap soap;
...
soap init(&soap);
soap set namespaces(&soap, namespaceTable1);
soap call remote method(&soap, URL, Action, ...);
...
10
gSOAP Serialization and Deserialization Rules
This section describes the serialization and deserialization of C and C++ data types for SOAP 1.1
and 1.2 compliant encoding and decoding.
10.1
SOAP RPC Encoding Versus Document/Literal and xsi:type Info
The wsdl2h tool automatically generates a header file specialized for SOAP RPC encoding or
document/literal style. The serialization and deserialization rules for C/C++ objects is almost
identical for these styles, except for the following important issues.
With SOAP RPC encoding style, care must be taken to ensure typed messages are produced
for interoperability and compatibility reasons. To ensure that the gSOAP engine automatically
generates typed (xsi:type attributed) messages, use soapcpp2 option -t, see also Section 8.1. While
gSOAP can handle untyped messages, some toolkits fail to find deserializers when the xsi:type
information is absent.
When starting the development of a gSOAP application from a header file, the soapcpp2 compiler
will generate WSDL and schema files for SOAP 1.1 document/literal style by default (use the
//gsoap directives to control this, see Section 18.2). Use soapcpp2 options -2, -e, and -t to generate
code for SOAP 1.2, RPC encoding, and typed messages.
98
With SOAP RPC encoding, generic complexTypes with maxOccurs="unbounded" are not allowed and
SOAP encoded arrays must be used. Also XML attributes and unions (XML schema choice) are
not allowed with SOAP RPC encoding.
Also with SOAP RPC encoding, multi-reference accessors are common to encode co-referenced
objects and object digraphs. Multi-reference encoding is not supported in document/literal style,
which means that cyclic object digraphs cannot be serialized (the engine will crash). Also DAGs
are represented as XML trees in document/literal style messaging.
10.2
Primitive Type Encoding
The default encoding rules for the primitive C and C++ data types are given in the table below:
Type
bool
char* (C string)
char
long double
double
float
int
long
LONG64
long long
short
time t
struct tm
unsigned char
unsigned int
unsigned long
ULONG64
unsigned long long
unsigned short
wchar t*
XSD Type
boolean
string
byte
decimal (with #import ”custom/long double.h”)
double
float
int
long
long
long
short
dateTime
dateTime (with #import ”custom/struct tm.h”)
unsignedByte
unsignedInt
unsignedLong
unsignedLong
unsignedLong
unsignedShort
string
Objects of type void and void* cannot be encoded. Enumerations and bit masks are supported as
well, see 10.4.
10.3
How to Encode and Decode Primitive Types as XSD Types
By default, encoding of the primitive types will take place as per SOAP encoding style. The
encoding can be changed to any XML Schema type (XSD type) with an optional namespace prefix
by using a typedef in the header file input to the gSOAP stub and skeleton compiler. The declaration
enables the implementation of built-in XML Schema types (also known as XSD types) such as
positiveInteger, xsd:anyURI, and xsd:date for which no built-in data structures in C and C++
exist but which can be represented using standard data structures such as strings, integers, and
floats.
99
The typedef declaration is frequently used for convenience in C. A typedef declares a type name
for a (complex) type expression. The type name can then be used in other declarations in place of
the more complex type expression, which often improves the readability of the program code.
The gSOAP compiler interprets typedef declarations the same way as a regular C compiler interprets
them, i.e. as types in declarations. In addition however, the gSOAP compiler will also use the type
name in the encoding of the data in SOAP. The typedef name will appear as the XML element
name of an independent element and as the value of the xsi:type attribute in the SOAP payload.
Many built-in primitive and derived XSD types such as xsd:anyURI, positiveInteger, and decimal
can be stored by standard primitive data structures in C++ as well such as strings, integers, floats,
and doubles. To serialize strings, integers, floats, and doubles as built-in primitive and derived
XSD types. To this end, a typedef declaration can be used to declare an XSD type.
For example, the declaration
typedef unsigned int xsd positiveInteger;
creates a named type positiveInteger which is represented by unsigned int in C++. For example, the
encoding of a positiveInteger value 3 is
<positiveInteger xsi:type="xsd:positiveInteger">3</positiveInteger>
The built-in primitive and derived numerical XML Schema types are listed below together with
their recommended typedef declarations. Note that the SOAP encoding schemas for primitive types
are derived from the built-in XML Schema types, so SOAP ENC can be used as a namespace prefix
instead of xsd .
xsd:anyURI Represents a Uniform Resource Identifier Reference (URI). Each URI scheme imposes
specialized syntax rules for URIs in that scheme, including restrictions on the syntax of
allowed fragment identifiers. It is recommended to use strings to store xsd:anyURI XML
Schema types. The recommended type declaration is:
typedef char *xsd anyURI;
xsd:base64Binary Represents Base64-encoded arbitrary binary data. For using the xsd:base64Binary
XSD Schema type, the use of the base64Binary representation of a dynamic array is strongly
recommended, see Section 10.12. However, the type can also be declared as a string and the
encoding will be string-based:
typedef char *xsd base64Binary;
With this approach, it is solely the responsibility of the application to make sure the string
content is according to the Base64 Content-Transfer-Encoding defined in Section 6.8 of RFC
2045.
xsd:boolean For declaring an xsd:boolean XSD Schema type, the use of a bool is strongly recommended. If a pure C compiler is used that does not support the bool type, see Section 10.4.5.
The corresponding type declaration is:
typedef bool xsd boolean;
100
Type xsd boolean declares a Boolean (0 or 1), which is encoded as
<xsd:boolean xsi:type="xsd:boolean">...</xsd:boolean>
xsd:byte Represents a byte (-128...127). The corresponding type declaration is:
typedef char xsd byte;
Type xsd byte declares a byte which is encoded as
<xsd:byte xsi:type="xsd:byte">...</xsd:byte>
xsd:dateTime Represents a date and time. The lexical representation is according to the ISO
8601 extended format CCYY-MM-DDThh:mm:ss where ”CC” represents the century, ”YY”
the year, ”MM” the month and ”DD” the day, preceded by an optional leading ”-” sign to
indicate a negative number. If the sign is omitted, ”+” is assumed. The letter ”T” is the
date/time separator and ”hh”, ”mm”, ”ss” represent hour, minute and second respectively.
It is recommended to use the time t type to store xsd:dateTime XSD Schema types and the
type declaration is:
typedef time t xsd dateTime;
However, note that calendar times before the year 1902 or after the year 2037 cannot be
represented. Upon receiving a date outside this range, the time t value will be set to -1.
Strings (char*) can be used to store xsd:dateTime XSD Schema types. The type declaration
is:
typedef char *xsd dateTime;
In this case, it is up to the application to read and set the dateTime representation.
xsd:date Represents a date. The lexical representation for date is the reduced (right truncated)
lexical representation for dateTime: CCYY-MM-DD. It is recommended to use strings (char*)
to store xsd:date XSD Schema types. The type declaration is:
typedef char *xsd date;
xsd:decimal Represents arbitrary precision decimal numbers. It is recommended to use the double
type to store xsd:decimal XSD Schema types and the type declaration is:
typedef double xsd decimal;
Type xsd decimal declares a double floating point number which is encoded as
<xsd:double xsi:type="xsd:decimal">...</xsd:double>
xsd:double Corresponds to the IEEE double-precision 64-bit floating point type. The type decla-
ration is:
typedef double xsd double;
Type xsd double declares a double floating point number which is encoded as
101
<xsd:double xsi:type="xsd:double">...</xsd:double>
xsd:duration Represents a duration of time. The lexical representation for duration is the ISO
8601 extended format PnYn MnDTnH nMnS, where nY represents the number of years, nM
the number of months, nD the number of days, T is the date/time separator, nH the number
of hours, nM the number of minutes and nS the number of seconds. The number of seconds
can include decimal digits to arbitrary precision. It is recommended to use strings (char*) to
store xsd:duration XSD Schema types. The type declaration is:
typedef char *xsd duration;
xsd:float Corresponds to the IEEE single-precision 32-bit floating point type. The type declara-
tion is:
typedef float xsd float;
Type xsd float declares a floating point number which is encoded as
<xsd:float xsi:type="xsd:float">...</xsd:float>
xsd:hexBinary Represents arbitrary hex-encoded binary data. It has a lexical representation where
each binary octet is encoded as a character tuple, consisting of two hexadecimal digits ([09a-fA-F]) representing the octet code. For example, ”0FB7” is a hex encoding for the 16-bit
integer 4023 (whose binary representation is 111110110111. For using the xsd:hexBinary
XSD Schema type, the use of the hexBinary representation of a dynamic array is strongly
recommended, see Section 10.13. However, the type can also be declared as a string and the
encoding will be string-based:
typedef char *xsd hexBinary;
With this approach, it is solely the responsibility of the application to make sure the string
content consists of a sequence of octets.
xsd:int Corresponds to a 32-bit integer in the range -2147483648 to 2147483647. If the C++
compiler supports 32-bit int types, the type declaration can use the int type:
typedef int xsd int;
Otherwise, the C++ compiler supports 16-bit int types and the type declaration should use
the long type:
typedef long xsd int;
Type xsd int declares a 32-bit integer which is encoded as
<xsd:int xsi:type="xsd:int">...</xsd:int>
xsd:integer Corresponds to an unbounded integer. Since C++ does not support unbounded inte-
gers as a standard feature, the recommended type declaration is:
typedef long long xsd integer;
Type xsd integer declares a 64-bit integer which is encoded as an unbounded xsd:integer:
102
<xsd:integer xsi:type="xsd:integer">...</xsd:integer>
Another possibility is to use strings to represent unbounded integers and do the translation
in code.
xsd:long Corresponds to a 64-bit integer in the range -9223372036854775808 to 9223372036854775807.
The type declaration is:
typedef long long xsd long;
Or in Visual C++:
typedef LONG64 xsd long;
Type xsd long declares a 64-bit integer which is encoded as
<xsd:long xsi:type="xsd:long">...</xsd:long>
xsd:negativeInteger Corresponds to a negative unbounded integer (< 0). Since C++ does not
support unbounded integers as a standard feature, the recommended type declaration is:
typedef long long xsd negativeInteger;
Type xsd negativeInteger declares a 64-bit integer which is encoded as a xsd:negativeInteger:
<xsd:negativeInteger xsi:type="xsd:negativeInteger">...</xsd:negativeInteger>
Another possibility is to use strings to represent unbounded integers and do the translation
in code.
xsd:nonNegativeInteger Corresponds to a non-negative unbounded integer (> 0). Since C++ does
not support unbounded integers as a standard feature, the recommended type declaration is:
typedef unsigned long long xsd nonNegativeInteger;
Type xsd nonNegativeInteger declares a 64-bit unsigned integer which is encoded as a nonnegative unbounded xsd:nonNegativeInteger:
<xsd:nonNegativeInteger xsi:type="xsd:nonNegativeInteger">...</xsd:nonNegativeInteger>
Another possibility is to use strings to represent unbounded integers and do the translation
in code.
xsd:nonPositiveInteger Corresponds to a non-positive unbounded integer (≤ 0). Since C++ does
not support unbounded integers as a standard feature, the recommended type declaration is:
typedef long long xsd nonPositiveInteger;
Type xsd nonPositiveInteger declares a 64-bit integer which is encoded as a xsd:nonPositiveInteger:
<xsd:nonPositiveInteger xsi:type="xsd:nonPositiveInteger">...</xsd:nonPositiveInteger>
Another possibility is to use strings to represent unbounded integers and do the translation
in code.
103
xsd:normalizedString Represents normalized character strings. Normalized character strings do
not contain the carriage return (#xD), line feed (#xA) nor tab (#x9) characters. It is
recommended to use strings to store xsd:normalizeString XSD Schema types. The type
declaration is:
typedef char *xsd normalizedString;
Type xsd normalizedString declares a string type which is encoded as
<xsd:normalizedString xsi:type="xsd:normalizedString">...</xsd:normalizedString>
It is solely the responsibility of the application to make sure the strings do not contain carriage
return (#xD), line feed (#xA) and tab (#x9) characters.
xsd:positiveInteger Corresponds to a positive unbounded integer (≥ 0). Since C++ does not
support unbounded integers as a standard feature, the recommended type declaration is:
typedef unsigned long long xsd positiveInteger;
Type xsd positiveInteger declares a 64-bit unsigned integer which is encoded as a xsd:positiveInteger:
<xsd:positiveInteger xsi:type="xsd:positiveInteger">...</xsd:positiveInteger>
Another possibility is to use strings to represent unbounded integers and do the translation
in code.
xsd:short Corresponds to a 16-bit integer in the range -32768 to 32767. The type declaration is:
typedef short xsd short;
Type xsd short declares a short 16-bit integer which is encoded as
<xsd:short xsi:type="xsd:short">...</xsd:short>
xsd:string Represents character strings. The type declaration is:
typedef char *xsd string;
Type xsd string declares a string type which is encoded as
<xsd:string xsi:type="xsd:string">...</xsd:string>
The type declaration for wide character strings is:
typedef wchar t *xsd string;
Both type of strings can be used at the same time, but requires one typedef name to be
changed by appending an underscore which is invisible in XML. For example:
typedef wchar t *xsd string ;
xsd:time Represents a time. The lexical representation for time is the left truncated lexical rep-
resentation for dateTime: hh:mm:ss.sss with optional following time zone indicator. It is
recommended to use strings (char*) to store xsd:time XSD Schema types. The type declaration is:
104
typedef char *xsd time;
xsd:token Represents tokenized strings. Tokens are strings that do not contain the line feed (#xA)
nor tab (#x9) characters, that have no leading or trailing spaces (#x20) and that have no
internal sequences of two or more spaces. It is recommended to use strings to store xsd:token
XSD Schema types. The type declaration is:
typedef char *xsd token;
Type xsd token declares a string type which is encoded as
<xsd:token xsi:type="xsd:token">...</xsd:token>
It is solely the responsibility of the application to make sure the strings do not contain the
line feed (#xA) nor tab (#x9) characters, that have no leading or trailing spaces (#x20) and
that have no internal sequences of two or more spaces.
xsd:unsignedByte Corresponds to an 8-bit unsigned integer in the range 0 to 255. The type decla-
ration is:
typedef unsigned char xsd unsignedByte;
Type xsd unsignedByte declares a unsigned 8-bit integer which is encoded as
<xsd:unsignedByte xsi:type="xsd:unsignedByte">...</xsd:unsignedByte>
xsd:unsignedInt Corresponds to a 32-bit unsigned integer in the range 0 to 4294967295. If the
C++ compiler supports 32-bit int types, the type declaration can use the int type:
typedef unsigned int xsd unsignedInt;
Otherwise, the C++ compiler supports 16-bit int types and the type declaration should use
the long type:
typedef unsigned long xsd unsignedInt;
Type xsd unsignedInt declares an unsigned 32-bit integer which is encoded as
<xsd:unsignedInt xsi:type="xsd:unsignedInt">...</xsd:unsignedInt>
xsd:unsignedLong Corresponds to a 64-bit unsigned integer in the range 0 to 18446744073709551615.
The type declaration is:
typedef unsigned long long xsd unsignedLong;
Or in Visual C++:
typedef ULONG64 xsd unsignedLong;
Type xsd unsignedLong declares an unsigned 64-bit integer which is encoded as
<xsd:unsignedLong xsi:type="xsd:unsignedLong">...</xsd:unsignedLong>
xsd:unsignedShort Corresponds to a 16-bit unsigned integer in the range 0 to 65535. The type
declaration is:
105
typedef unsigned short xsd unsignedShort;
Type xsd unsginedShort declares an unsigned short 16-bit integer which is encoded as
<xsd:unsignedShort xsi:type="xsd:unsignedShort">...</xsd:unsignedShort>
Other XSD Schema types such as gYearMonth, gYear, gMonthDay, gDay, xsd:gMonth, QName, NOTATION,
etc., can be encoded similarly using a typedef declaration.
10.3.1
How to Use Multiple C/C++ Types for a Single Primitive XSD Type
Trailing underscores (see Section 9.3) can be used in the type name in a typedef to enable the
declaration of multiple storage formats for a single XML Schema type. For example, one part
of a C/C++ application’s data structure may use plain strings while another part may use wide
character strings. To enable this simultaneous use, declare:
typedef char *xsd string;
typedef wchar t *xsd string ;
Now, the xsd string and xsd string types will both be encoded and decoded as XML string types
and the use of trailing underscores allows multiple declarations for a single XML Schema type.
10.3.2
How to use Wrapper Classes to Specify Polymorphic Primitive Types
SOAP 1.1 supports polymorphic types, because XSD Schema types form a hierarchy. The root of
the hierarchy is called xsd:anyType (xsd:ur-type in the older 1999 schema). So, for example, an
array of xsd:anyType in SOAP may actually contain any mix of element types that are the derived
types of the root type. The use of polymorphic types is indicated by the WSDL and schema
descriptions of a Web service and can therefore be predicted/expected for each particular case.
On the one hand, the typedef construct provides a convenient way to associate C/C++ types with
XML Schema types and makes it easy to incorporate these types in a (legacy) C/C++ application.
However, on the other hand the typedef declarations cannot be used to support polymorphic XML
Schema types. Most SOAP clients and services do not use polymorphic types. In case they do,
the primitive polymorphic types can be declared as a hierarchy of C++ classes that can be used
simultaneously with the typedef declarations.
The general form of a primitive type declaration that is derived from a super type is:
class xsd type name: [public xsd super type name]
{ public: Type item;
[public:] [private] [protected:]
method1;
method2;
...
};
where Type is a primitive C type. The
item field MUST be the first field in this wrapper class.
For example, the XML Schema type hierarchy can be copied to C++ with the following declarations:
106
class xsd anyType { };
class xsd anySimpleType: public xsd anyType { };
typedef char *xsd anyURI;
class xsd anyURI : public xsd anySimpleType { public: xsd anyURI item; };
typedef bool xsd boolean;
class xsd boolean : public xsd anySimpleType { public: xsd boolean item; };
typedef char *xsd date;
class xsd date : public xsd anySimpleType { public: xsd date item; };
typedef time t xsd dateTime;
class xsd dateTime : public xsd anySimpleType { public: xsd dateTime item; };
typedef double xsd double;
class xsd double : public xsd anySimpleType { public: xsd double item; };
typedef char *xsd duration;
class xsd duration : public xsd anySimpleType { public: xsd duration item; };
typedef float xsd float;
class xsd float : public xsd anySimpleType { public: xsd float item; };
typedef char *xsd time;
class xsd time : public xsd anySimpleType { public: xsd time item; };
typedef char *xsd decimal;
class xsd decimal : public xsd anySimpleType { public: xsd decimal item; };
typedef char *xsd integer;
class xsd integer : public xsd decimal { public: xsd integer item; };
typedef LONG64 xsd long;
class xsd long : public xsd integer { public: xsd long item; };
typedef long xsd int;
class xsd int : public xsd long { public: xsd int item; };
typedef short xsd short;
class xsd short : public xsd int { public: xsd short item; };
typedef char xsd byte;
class xsd byte : public xsd short { public: xsd byte item; };
typedef char *xsd nonPositiveInteger;
class xsd nonPositiveInteger : public xsd integer { public: xsd nonPositiveInteger item; };
typedef char *xsd negativeInteger;
class xsd negativeInteger : public xsd nonPositiveInteger { public: xsd negativeInteger item;
};
typedef char *xsd nonNegativeInteger;
class xsd nonNegativeInteger : public xsd integer { public: xsd nonNegativeInteger item; };
typedef char *xsd positiveInteger;
class xsd positiveInteger : public xsd nonNegativeInteger { public: xsd positiveInteger item;
};
typedef ULONG64 xsd unsignedLong;
class xsd unsignedLong : public xsd nonNegativeInteger { public: xsd unsignedLong item;
};
typedef unsigned long xsd unsignedInt;
class xsd unsignedInt : public xsd unsginedLong { public: xsd unsignedInt item; };
typedef unsigned short xsd unsignedShort;
class xsd unsignedShort : public xsd unsignedInt { public: xsd unsignedShort item; };
typedef unsigned char xsd unsignedByte;
class xsd unsignedByte : public xsd unsignedShort { public: xsd unsignedByte item; };
typedef char *xsd string;
class xsd string : public xsd anySimpleType { public: xsd string item; };
107
typedef char *xsd normalizedString;
class xsd normalizedString : public xsd string { public: xsd normalizedString
typedef char *xsd token;
class xsd token : public xsd normalizedString { public: xsd token item; };
item; };
Note the use of the trailing underscores for the class names to distinguish the typedef type names
from the class names. Only the most frequently used built-in schema types are shown. It is also
allowed to include the xsd:base64Binary and xsd:hexBinary types in the hierarchy:
class xsd base64Binary: public xsd anySimpleType { public: unsigned char * ptr; int size;
};
class xsd hexBinary: public xsd anySimpleType { public: unsigned char * ptr; int size; };
See Sections 10.12 and 10.13.
Methods are allowed to be added to the classes above, such as constructors and getter/setter
methods, see Section 10.6.4.
Wrapper structs are supported as well, similar to wrapper classes. But they cannot be used to
implement polymorphism. Rather, the wrapper structs facilitate the use of XML attributes with a
primitive typed object, see 10.6.7.
10.3.3
XSD Schema Type Decoding Rules
The decoding rules for the primitive C and C++ data types is given in the table below:
108
Type
bool
char* (C string)
wchar t * (wide string)
Allows Decoding of
[xsd:]boolean
any type, see 10.3.5
any type, see 10.3.5
Precision Lost?
no
no
no
double
[xsd:]double
[xsd:]float
[xsd:]long
[xsd:]int
[xsd:]short
[xsd:]byte
[xsd:]unsignedLong
[xsd:]unsignedInt
[xsd:]unsignedShort
[xsd:]unsignedByte
[xsd:]decimal
[xsd:]integer
[xsd:]positiveInteger
[xsd:]negativeInteger
[xsd:]nonPositiveInteger
[xsd:]nonNegativeInteger
no
no
no
no
no
no
no
no
no
no
possibly
possibly
possibly
possibly
possibly
possibly
float
[xsd:]float
[xsd:]long
[xsd:]int
[xsd:]short
[xsd:]byte
[xsd:]unsignedLong
[xsd:]unsignedInt
[xsd:]unsignedShort
[xsd:]unsignedByte
[xsd:]decimal
[xsd:]integer
[xsd:]positiveInteger
[xsd:]negativeInteger
[xsd:]nonPositiveInteger
[xsd:]nonNegativeInteger
no
no
no
no
no
no
no
no
no
possibly
possibly
possibly
possibly
possibly
possibly
long long
[xsd:]long
[xsd:]int
[xsd:]short
[xsd:]byte
[xsd:]unsignedLong
[xsd:]unsignedInt
[xsd:]unsignedShort
[xsd:]unsignedByte
[xsd:]integer
[xsd:]positiveInteger
[xsd:]negativeInteger
[xsd:]nonPositiveInteger
[xsd:]nonNegativeInteger
no
no
no
no
possibly
no
no
no
possibly
possibly
possibly
possibly
possibly
109
Type
long
Allows Decoding of
[xsd:]long
[xsd:]int
[xsd:]short
[xsd:]byte
[xsd:]unsignedLong
[xsd:]unsignedInt
[xsd:]unsignedShort
[xsd:]unsignedByte
Precision Lost?
possibly, if long is 32 bit
no
no
no
possibly
no
no
no
int
[xsd:]int
[xsd:]short
[xsd:]byte
[xsd:]unsignedInt
[xsd:]unsignedShort
[xsd:]unsignedByte
no
no
no
possibly
no
no
short
[xsd:]short
[xsd:]byte
[xsd:]unsignedShort
[xsd:]unsignedByte
no
no
no
no
char
[xsd:]byte
[xsd:]unsignedByte
no
possibly
unsigned long long
[xsd:]unsignedLong
[xsd:]unsignedInt
[xsd:]unsignedShort
[xsd:]unsignedByte
[xsd:]positiveInteger
[xsd:]nonNegativeInteger
no
no
no
no
possibly
possibly
unsigned long
[xsd:]unsignedLong
[xsd:]unsignedInt
[xsd:]unsignedShort
[xsd:]unsignedByte
possibly, if long is 32 bit
no
no
no
unsigned int
[xsd:]unsignedInt
[xsd:]unsignedShort
[xsd:]unsignedByte
no
no
no
unsigned short
[xsd:]unsignedShort
[xsd:]unsignedByte
no
no
unsigned char
[xsd:]unsignedByte
no
time t
[xsd:]dateTime
no(?)
110
Due to limitations in representation of certain primitive C++ types, a possible loss of accuracy may
occur with the decoding of certain XSD Schema types as is indicated in the table. The table does
not indicate the possible loss of precision of floating point values due to the textual representation
of floating point values in SOAP.
All explicitly declared XSD Schema encoded primitive types adhere to the same decoding rules.
For example, the following declaration:
typedef unsigned long long xsd nonNegativeInteger;
enables the encoding and decoding of xsd:nonNegativeInteger XSD Schema types (although decoding takes place with a possible loss of precision). The declaration also allows decoding of
xsd:positiveInteger XSD Schema types, because of the storage as a unsigned long long data type.
10.3.4
Multi-Reference Strings
If more than one char pointer points to the same string, the string is encoded as a multi-reference
value. Consider for example
char *s = ”hello”, *t = s;
The s and t variables are assigned the same string, and when serialized, t refers to the content of s:
<string id="123" xsi:type="string">hello</string>
...
<string href="#123"/>
The example assumed that s and t are encoded as independent elements.
Note: the use of typedef to declare a string type such as xsd string will not affect the multi-reference
string encoding. However, strings declared with different typedef s will never be considered multireference even when they point to the same string. For example
typedef char *xsd string;
typedef char *xsd anyURI;
xsd anyURI *s = ”http://www.myservice.com”;
xsd string *t = s;
The variables s and t point to the same string, but since they are considered different types their
content will not be shared in the SOAP payload through a multi-referenced string.
10.3.5
“Smart String” Mixed-Content Decoding
The implementation of string decoding in gSOAP allows for mixed content decoding. If the SOAP
payload contains a complex data type in place of a string, the complex data type is decoded in the
string as plain XML text.
For example, suppose the getInfo remote method returns some detailed information. The remote
method is declared as:
111
// Contents of header file ”getInfo.h”:
getInfo(char *detail);
The proxy of the remote method is used by a client to request a piece of information and the service
responds with:
HTTP/1.1 200 OK
Content-Type: text/xml
Content-Length: nnn
<SOAP-ENV:Envelope xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:SOAP-ENC="http://schemas.xmlsoap.org/soap/encoding/"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:xsd="http://www.w3.org/2001/XMLSchema"
<SOAP-ENV:Body>
<getInfoResponse>
<detail>
<picture>Mona Lisa by <i>Leonardo da Vinci</i></picture>
</detail>
</getInfoResponse>
</SOAP-ENV:Body>
</SOAP-ENV:Envelope>
As a result of the mixed content decoding, the detail string contains “<picture>Mona Lisa by
<i>Leonardo da Vinci</i></picture>”.
10.3.6
STL Strings
gSOAP supports STL strings std::string and std::wstring. For example:
typedef std::string xsd string;
class ns myClass
{ public:
xsd string s; // serialized with xsi:type=”xsd:string”
std::string t; // serialized without xsi:type
...
};
Caution: Please avoid mixing std::string and C strings (char*) in the header file when using SOAP
1.1 encoding. The problem is that multi-referenced strings in SOAP encoded messages cannot be
assigned simultaneously to a std::string and a char* string.
10.3.7
Changing the Encoding Precision of float and double Types
The double encoding format is by default set to “%.18G” (see a manual on printf text formatting in
C), i.e. at most 18 digits of precision to limit a loss in accuracy. The float encoding format is by
default “%.9G”, i.e. at most 9 digits of precision.
The encoding format of a double type can be set by assigning a format string to soap.double format,
where soap is a variable that contains the current runtime environment. For example:
112
struct soap soap;
soap init(&soap); // sets double format = ”%.18G”
soap.double format = ”%e”; // redefine
which causes all doubles to be encoded in scientific notation. Likewise, the encoding format of a
float type can be set by assigning a format string to the static soap float format string variable. For
example:
struct soap soap;
soap init(&soap); // sets float format = ”%.9G”
soap.float format = ”%.4f”; // redefine
which causes all floats to be encoded with four digits precision.
Caution: The format strings are not automatically reset before or after SOAP communications.
An error in the format string may result in the incorrect encoding of floating point values.
10.3.8
INF, -INF, and NaN Values of float and double Types
The gSOAP runtime stdsoap2.cpp and header file stdsoap2.h support the marshalling of IEEE INF,
-INF, and NaN representations. Under certain circumstances this may break if the hardware and/or
C/C++ compiler does not support these representations. To remove the representations, remove
the inclusion of the <math.h> header file from the stdsoap2.h file. You can control the representations
as well, which are defined by the macros:
#define
#define
#define
#define
#define
#define
10.4
FLT NAN
FLT PINFTY
FLT NINFTY
DBL NAN
DBL PINFTY
DBL NINFTY
Enumeration Serialization
Enumerations are generally useful for the declaration of named integer-valued constants, also called
enumeration constants.
10.4.1
Serialization of Symbolic Enumeration Constants
The gSOAP stub and skeleton compiler encodes the constants of enumeration-typed variables in
symbolic form using the names of the constants when possible to comply to SOAP’s enumeration
encoding style. Consider for example the following enumeration of weekdays:
enum weekday {Mon, Tue, Wed, Thu, Fri, Sat, Sun};
The enumeration-constant Mon, for example, is encoded as
113
<weekday xsi:type="weekday">Mon</weekday>
The value of the xsi:type attribute is the enumeration-type identifier’s name. If the element is
independent as in the example above, the element name is the enumeration-type identifier’s name.
The encoding of complex types such as enumerations requires a reference to an XML Schema
through the use of a namespace prefix. The namespace prefix can be specified as part of the
enumeration-type identifier’s name, with the usual namespace prefix conventions for identifiers.
This can be used to explicitly specify the encoding style. For example:
enum ns1 weekday {Mon, Tue, Wed, Thu, Fri, Sat, Sun};
The enumeration-constant Sat, for example, is encoded as:
<ns1:weekday xsi:type="ns1:weekday">Sat</ns1:weekday>
The corresponding XML Schema for this enumeration data type would be:
<xsd:element name="weekday" type="tns:weekday"/>
<xsd:simpleType name="weekday">
<xsd:restriction base="xsd:string">
<xsd:enumeration value="Mon"/>
<xsd:enumeration value="Tue"/>
<xsd:enumeration value="Wed"/>
<xsd:enumeration value="Thu"/>
<xsd:enumeration value="Fri"/>
<xsd:enumeration value="Sat"/>
<xsd:enumeration value="Sun"/>
</xsd:restriction>
</xsd:simpleType>
10.4.2
Encoding of Enumeration Constants
If the value of an enumeration-typed variable has no corresponding named constant, the value is
encoded as a signed integer literal. For example, the following declaration of a workday enumeration
type lacks named constants for Saturday and Sunday:
enum ns1 workday {Mon, Tue, Wed, Thu, Fri};
If the constant 5 (Saturday) or 6 (Sunday) is assigned to a variable of the workday enumeration type,
the variable will be encoded with the integer literals 5 and 6, respectively. For example:
<ns1:workday xsi:type="ns1:workday">5</ns1:workday>
Since this is legal in C++ and SOAP allows enumeration constants to be integer literals, this
method ensures that non-symbolic enumeration constants are correctly communicated to another
party if the other party accepts literal enumeration constants (as with the gSOAP stub and skeleton
compiler).
114
Both symbolic and literal enumeration constants can be decoded.
To enforce the literal enumeration constant encoding and to get the literal constants in the WSDL
file, use the following trick:
enum ns1 nums { 1 = 1, 2 = 2, 3 = 3 };
The difference with an enumeration type without a list of values and the enumeration type above
is that the enumeration constants will appear in the WSDL service description.
10.4.3
Initialized Enumeration Constants
The gSOAP compiler supports the initialization of enumeration constants, as in:
enum ns1 relation {LESS = -1, EQUAL = 0, GREATER = 1};
The symbolic names LESS, EQUAL, and GREATER will appear in the SOAP payload for the encoding
of the ns1 relation enumeration values -1, 0, and 1, respectively.
10.4.4
How to “Reuse” Symbolic Enumeration Constants
A well-known deficiency of C and C++ enumeration types is the lack of support for the reuse
of symbolic names by multiple enumerations. That is, the names of all the symbolic constants
defined by an enumeration cannot be reused by another enumeration. To force encoding of the
same symbolic name by different enumerations, the identifier of the symbolic name can end in an
underscore ( ) or any number of underscores to distinguish it from other symbolic names in C++.
This guarantees that the SOAP encoding will use the same name, while the symbolic names can
be distinguished in C++. Effectively, the underscores are removed from a symbolic name prior to
encoding.
Consider for example:
enum ns1 workday {Mon, Tue, Wed, Thu, Fri};
enum ns1 weekday {Mon , Tue , Wed , Thu , Fri , Sat , Sun };
which will result in the encoding of the constants of enum ns1 weekday without the underscore, for
example as Mon.
Caution: The following declaration:
enum ns1 workday {Mon, Tue, Wed, Thu, Fri};
enum ns1 weekday {Sat = 5, Sun = 6};
will not properly encode the weekday enumeration, because it lacks the named constants for workday
in its enumeration list.
115
10.4.5
Boolean Enumeration Serialization for C
When developing a C Web service application, the C++ bool type should not be used since it is
not usually supported by the C compiler. Instead, an enumeration type should be used to serialize
true/false values as xsd:boolean Schema type enumeration values. The xsd:boolean XML Schema
type is defined as:
enum xsd boolean {false , true };
The value false , for example, is encoded as:
<xsd:boolean xsi:type="xsd:boolean">false</xsd:boolean>
Peculiar of the SOAP boolean type encoding is that it only defines the values 0 and 1, while
the built-in XML Schema boolean type also defines the false and true symbolic constants as valid
values. The following example declaration of an enumeration type lacks named constants altogether
to force encoding of the enumeration values as literal constants:
enum SOAP ENC boolean {};
The value 0, for example, is encoded with an integer literal:
<SOAP-ENC:boolean xsi:type="SOAP-ENC:boolean">0<SOAP-ENC:boolean>
10.4.6
Bitmask Enumeration Serialization
A bitmask is an enumeration of flags such as declared with C#’s [Flags] enum annotation. gSOAP
supports bitmask encoding and decoding for interoperability. However, bitmask types are not
standardized with SOAP RPC.
A special syntactic convention is used in the header file input to the gSOAP compiler to indicate
the use of bitmasks with an asterisk:
enum * name { enum-constant, enum-constant, ... };
The gSOAP compiler will encode the enumeration constants as flags, i.e. as a series of powers of
2 starting with 1. The enumeration constants can be or-ed to form a bitvector (bitmask) which is
encoded and decoded as a list of symbolic values in SOAP. For example:
enum * ns machineStatus { ON, BELT, VALVE, HATCH};
int ns getMachineStatus(char *name, char *enum ns machineStatus result);
Note that the use of the enum does not require the asterisk, only the definition. The gSOAP
compiler generates the enumeration:
enum ns machineStatus { ON=1, BELT=2, VALVE=4, HATCH=8};
116
A remote method implementation in a Web service can return:
int ns getMachineStatus(struct soap *soap, char *name, enum ns machineStatus result)
{ ...
*result = BELT — HATCH;
return SOAP OK;
}
10.5
Struct Serialization
A struct data type is encoded as an XML Schema complexType (SOAP-encoded compound data
type) such that the struct name forms the data type’s element name and schema type and the fields
of the struct are the data type’s accessors. This encoding is identical to the class instance encoding
without inheritance and method declarations, see Section 10.6 for further details. However, the
encoding and decoding of structs is more efficient compared to class instances due to the lack of
inheritance and the requirement by the serialization routines to check inheritance properties at run
time.
Certain type of fields of a struct can be (de)serialized as XML attributes. See 10.6.7 for more
details.
10.6
Class Instance Serialization
A class instance is serialized as an XML Schema complexType (SOAP-encoded compound data
type) such that the class name forms the data type’s element name and schema type and the data
member fields are the data type’s accessors. Only the data member fields are encoded in the SOAP
payload. Class methods are not encoded.
The general form of a class declaration is:
class [namespace prefix ]class name1 [:[public:] [private:] [protected:] [namespace prefix ]class name2]
{
[public:] [private:] [protected:]
field1;
field2;
...
[public:] [private:] [protected:]
method1;
method2;
...
};
where
is the optional namespace prefix of the compound data type (see identifier translation rules 9.3)
namespace prefix
class name1 is the element name of the compound data type (see identifier translation rules 9.3).
117
class name2 is an optional base class.
field is a field declaration (data member). A field MAY be declared static and const and MAY be
initialized.
method is a method declaration. A method MAY be declared virtual, but abstract methods are not
allowed. The method parameter declarations are REQUIRED to have parameter identifier
names.
[public:] [private:] [protected:] are OPTIONAL. Only members with public acces permission
can be serialized.
A class name is REQUIRED to be unique and cannot have the same name as a struct, enum, or
remote method name specified in the header file input to the gSOAP compiler. The reason is that
remote method requests are encoded similarly to class instances in SOAP and they are in principle
undistinguishable (the method parameters are encoded just as the fields of a class).
Only single inheritance is supported by the gSOAP compiler. Multiple inheritance is not supported,
because of the limitations of the SOAP protocol.
If a constructor method is present, there MUST also be a constructor declaration with empty
parameter list.
Classes should be declared “volatile” if you don’t want gSOAP to add serialization methods to
these classes, see Section 18.4 for more details.
Class templates are not supported by the gSOAP compiler, but you can use STL containers, see
Section 10.11.8. You can also define your own containers similar to STL containers.
Certain fields of a class can be (de)serialized as XML attributes. See 10.6.7 for more details.
Arrays may be embedded within a class (and struct) using a pointer field and size information, see
Section 10.11.7. This defines what is sometimes referred to in SOAP as “generics”.
Void pointers may be used in a class (or struct), but you have to add a type field so the gSOAP
runtime can determine the type of object pointed to, see Section 10.9.
A class instance is encoded as:
<[namespace-prefix:]class-name xsi:type="[namespace-prefix:]class-name">
<basefield-name1 xsi:type="...">...</basefield-name1>
<basefield-name2 xsi:type="...">...</basefield-name2>
...
<field-name1 xsi:type="...">...</field-name1>
<field-name2 xsi:type="...">...</field-name2>
...
</[namespace-prefix:]class-name>
where the field-name accessors have element-name representations of the class fields and the
basefield-name accessors have element-name representations of the base class fields. (The optional
parts resulting from the specification are shown enclosed in [].)
The decoding of a class instance allows any ordering of the accessors in the SOAP payload. However,
if a base class field name is identical to a derived class field name because the field is overloaded,
118
the base class field name MUST precede the derived class field name in the SOAP payload for
decoding. gSOAP guarantees this, but interoperability with other SOAP implementations is cannot
be guaranteed.
10.6.1
Example
The following example declares a base class ns Object and a derived class ns Shape:
// Contents of file ”shape.h”:
class ns Object
{
public:
char *name;
};
class ns Shape : public ns Object
{
public:
int sides;
enum ns Color {Red, Green, Blue} color;
ns Shape();
ns Shape(int sides, enum ns Green color);
˜ns Shape();
};
The implementation of the methods of class ns Shape must not be part of the header file and need
to be defined elsewhere.
An instance of class ns Shape with name Triangle, 3 sides, and color Green is encoded as:
<ns:Shape xsi:type="ns:Shape">
<name xsi:type="string">Triangle</name>
<sides xsi:type="int">3</sides>
<color xsi:type="ns:Color">Green</color>
</ns:shape>
The namespace URI of the namespace prefix ns must be defined by a namespace mapping table,
see Section 9.4.
10.6.2
Initialized static const Fields
A data member field of a class declared as static const is initialized with a constant value at compile
time. This field is encoded in the serialization process, but is not decoded in the deserialization
process. For example:
// Contents of file ”triangle.h”:
class ns Triangle : public ns Object
{
public:
int size;
119
static const int sides = 3;
};
An instance of class ns Triangle is encoded in SOAP as:
<ns:Triangle xsi:type="ns:Triangle">
<name xsi:type="string">Triangle</name>
<size xsi:type="int">15</size>
<sides xsi:type="int">3>/sides>
</ns:Triangle>
Decoding will ignore the sides field’s value.
Caution: The current gSOAP implementation does not support encoding static const fields, due
to C++ compiler compatibility differences. This feature may be provided the future.
10.6.3
Class Methods
A class declaration in the header file input to the gSOAP compiler MAY include method declarations. The method implementations MUST NOT be part of the header file but are required to be
defined in another C++ source that is externally linked with the application. This convention is
also used for the constructors and destructors of the class.
Dynamic binding is supported, so a method MAY be declared virtual.
10.6.4
Getter and Setter Methods
Setter and getter methods are invoked at run time upon serialization and deserialization of class
instances, respectively. The use of setter and getter methods adds more flexibility to the serialization
and deserialization process.
A setter method is called in the serialization phase from the virtual soap serialization method generated by the gSOAP compiler. You can use setter methods to process a class instance just before it
is serialized. A setter method can be used to convert application data, such as translating transient
application data into serializable data, for example. You can also use setter methods to retrieve
dynamic content and use it to update a class instance right before serialization. Remember setters
as ”set to serialize” operations.
Getter methods are invoked after deserialization of the instance. You can use them to adjust the
contents of class instances after all their members have been deserialized. Getters can be used
to convert deserialized members into transient members and even invoke methods to process the
deserialized data on the fly.
Getter and setter methods have the following signature:
[virtual] int get(struct soap *soap) [const];
[virtual] int set(struct soap *soap);
The active soap struct will be passed to the get and set methods. The methods should return
SOAP OK when successful. A setter method should prepare the contents of the class instance for
serialization. A getter method should process the instance after deserialization.
120
Here is an example of a base64 binary class:
class xsd base64Binary
{ public:
unsignedchar * ptr;
int size;
int get(struct soap *soap);
int set(struct soap *soap);
};
Suppose that the type and options members of the attachment should be set when the class is about
to be serialized. This can be accomplished with the set method from the information provided by
the ptr to the data and the soap struct passed to the set method (you can pass data via the
void*soap.user field).
The get method is invoked after the base64 data has been processed. You can use it for postprocessing purposes.
Here is another example. It defines a primitive update type. The class is a wrapper for the time t
type, see Section 10.3.2. Therefore, elements of this type contain xsd:dateType data.
class update
{ public:
time t item;
int set(struct soap *soap);
};
The setter method assigns the current time:
int update::set(struct soap *soap)
{
this-> item = time(NULL);
return SOAP OK;
}
Therefore, serialization results in the inclusion of a time stamp in XML.
Caution: a get method is invoked only when the XML element with its data was completely parsed.
The method is not invoked when the element is an xsi:nil element or has an href attribute.
Caution: The soap serialize method of a class calls the setter (when provided). However, the
soap serialize method is declared const while the setter should be allowed to modify the contents
of the class instance. Therefore, the gSOAP-generated code recasts the instance and the const is
removed when invoking the setter.
10.6.5
Streaming XML with Getter and Setter Methods
Getter methods enable streaming XML operations. A getter method is invoked when the object
is deserialized and the rest of the SOAP/XML message has not been processed yet. For example,
you can add a getter method to the SOAP Header class to implement header processing logic that
is activated as soon as the SOAP Header is received. An example code is shown below:
121
class h Authentication
{ public:
char *id;
int get(struct soap *soap);
};
class SOAP ENV Header
{ public:
h Authentication *h authentication;
};
The Authentication SOAP Header field is instantiated and decoded. After decoding, the getter
method is invoked, which can be used to check the id before the rest of the SOAP message is
processed.
10.6.6
Polymorphism, Derived Classes, and Dynamic Binding
Interoperability between client and service applications developed with gSOAP is established even
when clients and/or services use derived classes instead of the base classes used in the declaration
of the remote method parameters. A client application MAY use pointers to instances of derived
classes for the input parameters of a remote method. If the service was compiled with a declaration and implementation of the derived class, the remote method base class input parameters are
demarshalled and a derived class instance is created instead of a base class instance. If the service
did not include a declaration of the derived class, the derived class fields are ignored and a base
class instance is created. Therefore, interoperability is guaranteed even when the client sends an
instance of a derived classes and when a service returns an instance of a derived class.
The following example declares Base and Derived classes and a remote method that takes a pointer
to a Base class instance and returns a Base class instance:
// Contents of file ”derived.h”
class Base
{
public:
char *name;
Base();
virtual void print();
};
class Derived : public Base
{
public:
int num;
Derived();
virtual void print();
};
int method(Base *in, struct methodResponse { Base *out; } &result);
This header file specification is processed by the gSOAP compiler to produce the stub and skeleton
routines which are used to implement a client and service. The pointer of the remote method is
also allowed to point to Derived class instances and these instances will be marshalled as Derived
122
class instances and send to a service, which is in accord to the usual semantics of parameter passing
in C++ with dynamic binding.
The Base and Derived class method implementations are:
// Method implementations of the Base and Derived classes:
#include ”soapH.h”
...
Base::Base()
{
cout << ”created a Base class instance” << endl;
}
Derived::Derived()
{
cout << ”created a Derived class instance” << endl;
}
Base::print()
{
cout << ”print(): Base class instance ” << name << endl;
}
Derived::print()
{
cout << ”print(): Derived class instance ” << name << ” ” << num << endl;
}
Below is an example CLIENT application that creates a Derived class instance that is passed as the
input parameter of the remote method:
// CLIENT
#include ”soapH.h”
int main()
{
struct soap soap;
soap init(&soap);
Derived obj1;
Base *obj2;
struct methodResponse r;
obj1.name = ”X”;
obj1.num = 3;
soap call method(&soap, url, action, &obj1, r);
r.obj2->print();
}
...
The following example SERVER1 application copies a class instance (Base or Derived class) from
the input to the output parameter:
// SERVER1
#include ”soapH.h”
int main()
{
123
soap serve(soap new());
}
int method(struct soap *soap, Base *obj1, struct methodResponse &result)
{
obj1->print();
result.obj2 = obj1;
return SOAP OK;
}
...
The following messages are produced by the CLIENT and SERVER1 applications:
CLIENT: created a Derived class instance
SERVER1: created a Derived class instance
SERVER1: print(): Derived class instance X 3
CLIENT: created a Derived class instance
CLIENT: print(): Derived class instance X 3
Which indicates that the derived class kept its identity when it passed through SERVER1. Note
that instances are created both by the CLIENT and SERVER1 by the demarshalling process.
Now suppose a service application is developed that only accepts Base class instances. The header
file is:
// Contents of file ”base.h”:
class Base
{
public:
char *name;
Base();
virtual void print();
};
int method(Base *in, Base *out);
This header file specification is processed by the gSOAP stub and skeleton compiler to produce
skeleton routine which is used to implement a service (so the client will still use the derived classes).
The method implementation of the Base class are:
// Method implementations of the Base class:
#include ”soapH.h”
...
Base::Base()
{
cout << ”created a Base class instance” << endl;
}
Base::print()
{
cout << ”print(): Base class instance ” << name << endl;
}
124
And the SERVER2 application is that uses the Base class is:
// SERVER2
#include ”soapH.h”
int main()
{
soap serve(soap new());
}
int method(struct soap *soap, Base *obj1, struct methodResponse &result)
{
obj1->print();
result.obj2 = obj1;
return SOAP OK;
}
...
Here are the messages produced by the CLIENT and SERVER2 applications:
CLIENT: created a Derived class instance
SERVER2: created a Base class instance
SERVER2: print(): Base class instance X
CLIENT: created a Base class instance
CLIENT: print(): Base class instance X
In this example, the object was passed as a Derived class instance to SERVER2. Since SERVER2
only implements the Base class, this object is converted to a Base class instance and send back to
CLIENT.
10.6.7
XML Attributes
The SOAP RPC/LIT and SOAP DOC/LIT encoding styles support XML attributes in SOAP
messages while SOAP RPC with “Section 5” encoding does not support XML attributes other
than the SOAP and XSD specific attributes. SOAP RPC “Section 5” encoding has advantages for
cross-language interoperability and data encodings such as graph serialization. However, RPC/LIT
and DOC/LIT enables direct exchange of XML documents, which may include encoded application
data structures. Language interoperability is compromised, because no mapping between XML and
the typical language data types is defined. The meaning of the RPC/LIT and DOC/LIT XML
content is Schema driven rather than application/language driven.
gSOAP supports XML attribute (de)serialization of members in structs and classes. Attributes
are primitive XSD types, such as strings, enumerations, boolean, and numeric types. To declare
an XML attribute in a struct/class, the qualifier @ is used with the type of the attribute. The
type must be primitive type (including enumerations and strings), which can be declared with or
without a typedef to associate a XSD type with the C/C+ type. For example
typedef char *xsd string;
typedef bool *xsd boolean;
enum ns state { 0, 1, 2 };
struct ns myStruct
125
{
@ xsd string ns type; // encode as XML attribute ’ns:type’ of type ’xsd:string’
@ xsd boolean ns flag = false; // encode as XML attribute ’ns:flag’ of type ’xsd:boolean’
@ enum ns state ns state = 2; // encode as XML attribute ’ns:state’ of type ’ns:state’
struct ns myStruct *next;
};
The @ qualifier indicates XML attribute encoding for the ns type, ns flag, and ns state fields. Note
that the namespace prefix ns is used to distinguish these attributes from any other attributes such
as xsi:type (ns:type is not to be confused with xsi:type).
Default values can be associated with any field that has a primitive type in a struct/class, as is
illustrated in this example. The default values are used when the receiving message does not contain
the corresponding values.
String attributes are optional. Other type of attributes should be declared as pointers to make
them optional:
struct ns myStruct
{
@int *a; // omitted when NULL };
Because a remote method request and response is essentially a struct, XML attributes can also be
associated with method requests and responses. For example
int ns myMethod(@char *ns name, ...);
Attributes can also be attached to the dynamic arrays, binary types, and wrapper classes/structs
of primitive types. Wrapper classes are described in Section 10.3.2. For example
struct xsd string
{
char * item;
@ xsd boolean flag;
};
and
struct xsd base64Binary
{
unsigned char * ptr;
int size;
@ xsd boolean flag;
};
The attribute declarations MUST follow the item, ptr, and
teristics of wrapper structs/classes and dynamic arrays.
size fields which define the charac-
Caution: Do not use XML attributes with SOAP RPC encoding. You can only use attributes
with RPC literal encoding.
126
10.6.8
QName Attributes and Elements
gSOAP ensures the proper decoding of XSD QNames. An element or attribute with type QName
(Qualified Name) contains a namespace prefix and a local name. You can declare a QName type
as a typedef char *xsd QName. Values of type QName are internally handled as regular strings.
gSOAP takes care of the proper namespace prefix mappings when deserializing QName values. For
example
typedef char *xsd QName;
struct ns myStruct
{
xsd QName elt = ”ns:xyz”; // QName element with default value ”ns:xyz”
@ xsd QName att = ”ns:abc”; // QName attribute with default value ”ns:abc”
};
When the elt and att fields are serialized, their string contents are just transmitted (which means
that the application is responsible to ensure proper formatting of the QName strings prior to
transmission). When the fields are deserialized however, gSOAP takes care mapping the qualifiers
to the appropriate namespace prefixes. Suppose that the inbound value for the elt is x:def, where
the namespace name associated with the prefix x matches the namespace name of the prefix ns (as
defined in the namespace mapping table). Then, the value is automatically converted into ns:def.
If the namespace name is not in the table, then x:def is converted to ”URI”:def where "URI" is the
namespace URI bound to x in the message received. This enables an application to retrieve the
namespace information, whether it is in the namespace mapping table or not.
Note: QName is a pre-defined typedef type and used by gSOAP to (de)serialize SOAP Fault codes
which are QName elements.
10.7
Union Serialization
A union is only serialized if the union is used within a struct or class declaration that includes a int
union field that acts as a discriminant or selector for the union fields. The selector stores run-time
usage information about the union fields. That is, the selector is used to enumerate the union fields
such that the gSOAP engine is able to select the correct union field to serialize.
A union within a struct or class with a selector field represents xs:choice within a Schema complexType component. For example:
struct ns PO
{ ... };
struct ns Invoice
{ ... };
union ns PO or Invoice
{
struct ns PO po;
struct ns Invoice invoice;
}; struct ns composite
{
char *name;
127
int union;
union ns PO or Invoice value;
};
The union ns PO or Invoice is expanded as a xs:choice:
<complexType name="composite">
<sequence>
<element name="name" type="xsd:string"/>
<choice>
<element name="po" type="ns:PO"/>
<element name="invoice" type="ns:Invoice"/>
</choice>
</sequence>
</complexType>
Therefore, the name of the union and field can be freely chosen. However, the union name should be
qualified (as shown in the example) to ensure instances of XML Schemas with elementFormDefault="qualified"
are correctly serialized (po and invoice are ns: qualified).
The int union field selector’s values are determined by the soapcpp2 compiler. Each union field
name has a selector value formed by:
SOAP UNION union-name field-name
These selector values enumerate the union fields starting with 1. The value 0 can be assigned to
omit the serialization of the union, but only if explicitly allowed by validation rules, which requires
minOccurs="0" for the xs:choice as follows:
struct ns composite
{
char *name;
int union 0; // <choice minOccurs="0">
union ns PO or Invoice value;
};
This way we can treat the union as an optional data item by setting
union=0.
The following example shows how the struct ns composite instance is initialized for serialization:
struct ns composite data;
data.name = ”...”;
data. union = SOAP UNION ns PO or Invoice po; // select PO
data.value.po.number = ...; // populate the PO
Note that failing to set the selector to a valid union field can lead to a crash of the gSOAP serializer
because it will attempt to serialize an invalid union field.
For deserialization of union types, the union selector will be ser to 0 (when permitted) by the
gSOAP deserializer or set to one of the union field selector values as determined by the XML
payload.
128
When more than one union is used in a struct or class, the
avoid name clashes by using suffixes as in:
union selectors must be renamed to
struct ns composite
{
char *name;
int union value; // added suffix ” value”
union ns PO or Invoice value;
int union data; // added suffix ” data”
union ns Email or Fax data;
};
10.8
Serializing Pointer Types
The serialization of a pointer to a data type amounts to the serialization of the data type in SOAP
and the SOAP encoded representation of a pointer to the data type is indistinguishable from the
encoded representation of the data type pointed to.
10.8.1
Multi-Referenced Data
A data structure pointed to by more than one pointer is serialized as SOAP multi-reference data.
This means that the data will be serialized only once and identified with a unique id attribute. The
encoding of the pointers to the shared data is done through the use of href attributes to refer to the
multi-reference data (also see Section 8.12 on options to control the serialization of multi-reference
data). Cyclic C/C++ data structures are encoded with multi-reference SOAP encoding. Consider
for example the following a linked list data structure:
typedef char *xsd string;
struct ns list
{
xsd string value;
struct ns list *next;
};
Suppose a cyclic linked list is created. The first node contains the value ”abc” and points to a node
with value ”def” which in turn points to the first node. This is encoded as:
<ns:list id="1" xsi:type="ns:list">
<value xsi:type="xsd:string">abc</value>
<next xsi:type="ns:list">
<value xsi:type="xsd:string">def</value>
<next href="#1"/>
</next>
</ns:list>
In case multi-referenced data is received that “does not fit in a pointer-based structure”, the data is
copied. For example, the following two structs are similar, except that the first uses pointer-based
fields while the other uses non-pointer-based fields:
129
typedef long xsd int;
struct ns record
{
xsd int *a;
xsd int *b;
} P;
struct ns record
{
xsd int a;
xsd int b;
} R;
...
P.a = &n;
P.b = &n;
...
Since both a and b fields of P point to the same integer, the encoding of P is multi-reference:
<ns:record xsi:type="ns:record">
<a href="#1"/>
<b href="#1"/>
</ns:record>
<id id="1" xsi:type="xsd:int">123</id>
Now, the decoding of the content in the R data structure that does not use pointers to integers
results in a copy of each multi-reference integer. Note that the two structs resemble the same XML
data type because the trailing underscore will be ignored in XML encoding and decoding.
10.8.2
NULL Pointers and Nil Elements
A NULL pointer is not serialized, unless the pointer itself is pointed to by another pointer (but see
Section 8.12 to control the serialization of NULLs). For example:
struct X
{
int *p;
int **q;
}
Suppose pointer q points to pointer p and suppose p=NULL. In that case the p pointer is serialized
as
<...
id="123" xsi:nil="true"/>
and the serialization of q refers to href="#123". Note that SOAP 1.1 does not support pointer to
pointer types (!), so this encoding is specific to gSOAP. The pointer to pointer encoding is rarely
used in codes anyway. More common is a pointer to a data type such as a struct with pointer fields.
Caution: When the deserializer encounters an XML element that has a xsi:nil="true" attribute
but the corresponding C++ data is not a pointer or reference, the deserializer will terminate with a
130
SOAP NULL fault when the SOAP XML STRICT flag is set. The types section of a WSDL description
contains information on the “nilability” of data.
10.9
Void Pointers
In general, void pointers (void*) cannot be (de)serialized because the type of data referred to is
untyped. To enable the (de)serialization of the void pointers that are members of structs or classes,
you can insert a int type field right before the void pointer field. The int type field contains run
time information on the type of the data pointed to by void* member in a struct/class to enable
the (de)serialization of this data. The int type field is set to a SOAP TYPE X value, where X is the
name of a type. gSOAP generates the SOAP TYPE X definitions in soapH.h and uses them internally
to uniquely identify the type of each object. The type naming conventions outlined in Section 7.5.1
are used to determine the type name for X.
Here is an example to illustrate the (de)serialization of a void* field in a struct:
struct myStruct
{
int type; // the SOAP TYPE pointed to by p
void *p;
};
The
type integer can be set to 0 at run time to omit the serialization of the void pointer field.
The following example illustrates the initialization of myStruct with a void pointer to an int:
struct myStruct S;
int n;
S.p = &n;
S. type = SOAP TYPE int;
The serialized output of S contains the integer.
The deserializer for myStruct will automatically set the type field and void pointer to the deserialized
data, provided that the XML content for p carries the xsi:type attribute from which gSOAP can
determine the type.
Important: when (de)serializing strings via a void* field, the void* pointer MUST directly point
to the string value rather than indirectly as with all other types. For example:
struct myStruct S;
S.p = (void*)”Hello”;
S. type = SOAP TYPE string;
This is the case for all string-based types, including types defined with typedef char*.
You may use an arbitrary suffix with the
For example
type fields to handle multiple void pointers in structs/classes.
131
struct myStruct
{
int typeOfp; // the SOAP TYPE pointed to by p
void *p;
int typeOfq; // the SOAP TYPE pointed to by q
void *q;
};
Because service method parameters are stored within structs, you can use type and void* parameters to pass polymorphic arguments without having to define a C++ class hierarchy (Section 10.6.6).
For example:
typedef char *xsd string;
typedef int xsd int;
typedef float xsd float;
enum ns status { on, off };
struct ns widget { xsd string name; xsd int part; }; int ns myMethod(int
struct ns myMethodResponse { int type; void *return ; } *out);
type, void *data,
This method has a polymorphic input parameter data and a polymorphic output parameter return .
The type parameters can be one of SOAP TYPE xsd string, SOAP TYPE xsd int, SOAP TYPE xsd float,
SOAP TYPE ns status, or SOAP TYPE ns widget. The WSDL produced by the gSOAP compiler
declares the polymorphic parameters of type xsd:anyType which is ”too loose” and doesn’t allow the gSOAP importer to handle the WSDL accurately. Future gSOAP releases might replace
xsd:anyType with a choice schema type that limits the choice of types to the types declared in the
header file.
10.10
Fixed-Size Arrays
Fixed size arrays are encoded as per SOAP 1.1 one-dimensional array types. Multi-dimensional
fixed size arrays are encoded by gSOAP as nested one-dimensional arrays in SOAP. Encoding of
fixed size arrays supports partially transmitted and sparse array SOAP formats.
The decoding of (multi-dimensional) fixed-size arrays supports the SOAP multi-dimensional array
format as well as partially transmitted and sparse array formats.
An example:
// Contents of header file ”fixed.h”:
struct Example
{
float a[2][3];
};
This specifies a fixed-size array part of the struct Example. The encoding of array a is:
<a xsi:type="SOAP-ENC:Array" SOAP-ENC:arrayType="float[][2]">
<SOAP-ENC:Array xsi:type="SOAP-ENC:Array" SOAP-ENC:arrayType="float[3]"
<float xsi:type="float">...</float>
132
<float xsi:type="float">...</float>
<float xsi:type="float">...</float>
</SOAP-ENC:Array>
<SOAP-ENC:Array xsi:type="SOAP-ENC:Array" SOAP-ENC:arrayType="float[3]"
<float xsi:type="float">...</float>
<float xsi:type="float">...</float>
<float xsi:type="float">...</float>
</SOAP-ENC:Array>
</a>
Caution: Any decoded parts of a (multi-dimensional) array that do not “fit” in the fixed size array
are ignored by the deserializer.
10.11
Dynamic Arrays
As the name suggests, dynamic arrays are much more flexible than fixed-size arrays and dynamic
arrays are better adaptable to the SOAP encoding and decoding rules for arrays. In addition,
a typical C application allocates a dynamic array using malloc, assigns the location to a pointer
variable, and deallocates the array later with free. A typical C++ application allocates a dynamic
array using new, assigns the location to a pointer variable, and deallocates the array later with
delete. Such dynamic allocations are flexible, but pose a problem for the serialization of data: how
does the array serializer know the length of the array to be serialized given only a pointer to the
sequence of elements? The application stores the size information somewhere. This information
is crucial for the array serializer and has to be made explicitly known to the array serializer by
packaging the pointer and array size information within a struct or class.
10.11.1
SOAP Array Bounds Limits
SOAP encoded arrays use the SOAP-ENC:Array type and the SOAP-ENC:arrayType attribute to define
the array dimensionality and size. As a security measure to avoid denial of service attacks based on
sending a huge array size value requiring the allocation of large chunks of memory, the total number
of array elements set by the SOAP-ENC:arrayType attribute cannot exceed SOAP MAXARRAYSIZE,
which is set to 100,000 by default. This constant is defined in stdsoap2.h. This constant only
affects multi-dimensional arrays and the dimensionality of the receiving array will be lost when the
number of elements exceeds 100,000. One-dimensional arrays will be populated in sequential order
as expected.
10.11.2
One-Dimensional Dynamic Arrays
A special form of struct or class is used for one-dimensional dynamic arrays that contains a pointer
variable and a field that records the number of elements the pointer points to in memory.
The general form of the struct declaration for one-dimensional dynamic SOAP arrays is:
struct some name
{
Type * ptr; // pointer to array
133
int size; // number of elements pointed to
[[static const] int offset [= ...];] // optional SOAP 1.1 array offset
... // anything that follows here will be ignored
};
where Type MUST be a type associated with an XML Schema or MUST be a primitive type. If
these conditions are not met, a vector-like XML (de)serialization is used (see Section 10.11.7). A
primitive type can be used with or without a typedef . If the array elements are structs or classes,
then the struct/class type names should have a namespace prefix for schema association, or they
should be other (nested) dynamic arrays.
An alternative to a struct is to use a class with optional methods that MUST appear after the
and size fields:
ptr
class some name
{
public:
Type * ptr;
int size;
[[static const] int offset [= ...];]
method1;
method2;
... // any fields that follow will be ignored
};
To encode the data type as an array, the name of the struct or class SHOULD NOT have a namespace
prefix, otherwise the data type will be encoded and decoded as a generic vector, see Section 10.11.7.
The deserializer of a dynamic array can decode partially transmitted and/or SOAP sparse arrays,
and even multi-dimensional arrays which will be collapsed into a one-dimensional array with rowmajor ordering.
Caution: SOAP 1.2 does not support partially transmitted arrays. So the
array is ignored.
10.11.3
offset field of a dynamic
Example
The following example header file specifies the XMethods Service Listing service getAllSOAPServices
remote method and an array of SOAPService data structures:
// Contents of file ”listing.h”:
class ns3 SOAPService
{
public:
int ID;
char *name;
char *owner;
char *description;
char *homepageURL;
char *endpoint;
134
char
char
char
char
char
char
char
char
char
char
*SOAPAction;
*methodNamespaceURI;
*serviceStatus;
*methodName;
*dateCreated;
*downloadURL;
*wsdlURL;
*instructions;
*contactEmail;
*serverImplementation;
};
class ServiceArray
{
public:
ns3 SOAPService * ptr; // points to array elements
int size; // number of elements pointed to
ServiceArray();
˜ServiceArray();
void print();
};
int ns getAllSOAPServices(ServiceArray &return );
An example client application:
#include ”soapH.h” ...
// ServiceArray class method implementations:
ServiceArray::ServiceArray()
{
ptr = NULL;
size = 0;
}
ServiceArray::˜ServiceArray()
{ // destruction handled by gSOAP
}
void ServiceArray::print()
{
for (int i = 0; i ¡ size; i++)
cout << ptr[i].name << ”: ” << ptr[i].homepage << endl;
}
...
// Request a service listing and display results:
{
struct soap soap;
ServiceArray result;
const char *endpoint = ”www.xmethods.net:80/soap/servlet/rpcrouter”;
const char *action = ”urn:xmethodsServicesManager#getAllSOAPServices”;
...
soap init(&soap);
soap call ns getAllSOAPServices(&soap, endpoint, action, result);
result.print();
...
135
soap destroy(&soap); // dealloc class instances
soap end(&soap); // dealloc deserialized data
soap done(&soap); // cleanup and detach soap struct
}
10.11.4
One-Dimensional Dynamic Arrays With Non-Zero Offset
The declaration of a dynamic array as described in 10.11 MAY include an int offset field. When
set to an integer value, the serializer of the dynamic array will use this field as the start index of
the array and the SOAP array offset attribute will be used in the SOAP payload. Note that array
offsets is a SOAP 1.1 specific feature which is not supported in SOAP 1.2.
For example, the following header file declares a mathematical Vector class, which is a dynamic
array of floating point values with an index that starts at 1:
// Contents of file ”vector.h”:
typedef float xsd float;
class Vector
{
xsd float * ptr;
int size;
int offset;
Vector();
Vector(int n);
float& operator[](int i);
}
The implementations of the Vector methods are:
Vector::Vector()
{
ptr = NULL;
size = 0;
offset = 1;
}
Vector::Vector(int n)
{
ptr = (float*)malloc(n*sizeof(float));
size = n;
offset = 1;
}
Vector::˜Vector()
{
if ( ptr)
free( ptr);
}
float& Vector::operator[](int i)
{
return ptr[i- offset];
}
136
An example program fragment that serializes a vector of 3 elements:
struct soap soap;
soap init(&soap);
Vector v(3);
v[1] = 1.0;
v[2] = 2.0;
v[3] = 3.0;
soap begin(&soap);
v.serialize(&soap);
v.put("vec");
soap end(&soap);
The output is a partially transmitted array:
<vec xsi:type="SOAP-ENC:Array" SOAP-ENC:arrayType="xsd:float[4]" SOAP-ENC:offset="[1]">
<item xsi:type="xsd:float">1.0</item>
<item xsi:type="xsd:float">2.0</item>
<item xsi:type="xsd:float">3.0</item>
</vec>
Note that the size of the encoded array is necessarily set to 4 and that the encoding omits the
non-existent element at index 0.
The decoding of a dynamic array with an offset field is more efficient than decoding a dynamic array without an offset field, because the offset field will be assigned the value of the
SOAP-ENC:offset attribute instead of padding the initial part of the array with default values.
10.11.5
Nested One-Dimensional Dynamic Arrays
One-dimensional dynamic arrays MAY be nested. For example, using class Vector declared in the
previous section, class Matrix is declared:
// Contents of file ”matrix.h”:
class Matrix
{
public:
Vector * ptr;
int size;
int offset;
Matrix();
Matrix(int n, int m);
˜Matrix();
Vector& operator[](int i);
};
The Matrix type is essentially an array of pointers to arrays which make up the rows of a matrix.
The encoding of the two-dimensional dynamic array in SOAP will be in nested form.
137
10.11.6
Multi-Dimensional Dynamic Arrays
The general form of the struct declaration for K-dimensional (K > 1) dynamic arrays is:
struct some name
{
Type * ptr;
int size[K];
int offset[K];
... // anything that follows here will be ignored
};
where Type MUST be a type associated with an XML Schema, which means that it must be a
typedef ed type in case of a primitive type, or a struct/class name with a namespace prefix for
schema association, or another dynamic array. If these conditions are not met, a generic vector
XML (de)serialization is used (see Section 10.11.7).
An alternative is to use a class with optional methods:
class some name
{
public:
Type * ptr;
int size[K];
int offset[K];
method1;
method2;
... // any fields that follow will be ignored
};
In the above, K is a constant denoting the number of dimensions of the multi-dimensional array.
To encode the data type as an array, the name of the struct or class SHOULD NOT have a namespace
prefix, otherwise the data type will be encoded and decoded as a generic vector, see Section 10.11.7.
The deserializer of a dynamic array can decode partially transmitted multi-dimensional arrays.
For example, the following declaration specifies a matrix class:
typedef double xsd double;
class Matrix
{
public:
xsd double * ptr;
int size[2];
int offset[2];
};
In contrast to the matrix class of Section 10.11.5 that defined a matrix as an array of pointers
to matrix rows, this class has one pointer to a matrix stored in row-major order. The size of the
matrix is determined by the size field: size[0] holds the number of rows and size[1] holds the
number of columns of the matrix. Likewise, offset[0] is the row offset and offset[1] is the columns
offset.
138
10.11.7
Encoding XML Generics Containing Dynamic Arrays
XML “generics” extend the concept of a struct by allowing repetitions of elements within the struct.
A simple generic is an array-like data structure with a repetition of one element. To achieve this,
declare a dynamic array as a struct or class with a name that is qualified with a namespace prefix.
SOAP arrays are declared without prefix.
For example:
struct ns Map
{
struct ns Binding {char *key; char *val;} * ptr;
int size;
};
This declares a dynamic array, but the array will be serialized and deserialized as a generic with a
list-like data structure. For example:
<ns:Map xsi:type="ns:Map">
<ns:Binding xsi:type="ns:Binding">
<key>Joe</key>
<val>555 77 1234</val>
</ns:Binding>
<ns:Binding xsi:type="ns:Binding">
<key>Susan</key>
<val>555 12 6725</val>
</ns:Binding>
<ns:Binding xsi:type="ns:Binding">
<key>Pete</key>
<val>555 99 4321</val>
</ns:Binding>
</ns:Map>
Deserialization is less efficient compared to an array, because the size of the list is not part of the
SOAP encoding. Internal buffering is used by the deserializer to collect the elements. When the
end of the list is reached, the buffered elements are copied to a newly allocated space on the heap
for the dynamic array.
Multiple arrays can be used in a struct/class to support the concept of generics. Each array
results in a repetition of elements in the struct/class. This is achieved with a int size field in the
struct/class where the next field (i.e. below the size field) is a pointer type. The pointer type is
assumed to point to an array of values at run time. The size field holds the number of values at
run time. Multiple arrays can be embedded in a struct/class with size fields that have a distinct
names. To make the size fields distinct, you can end them with a unique name suffix such as
sizeOfstrings, for example.
The general convention for embedding arrays is:
struct ns SomeStruct
{
139
...
int sizename1; // number of elements pointed to
Type1 *field1; // by this field
...
int sizename2; // number of elements pointed to
Type2 *field2; // by this field
...
};
where name1 and name2 are identifiers used as a suffix to distinguish the
can be arbitrary and are not visible in XML.
size field. These names
For example, the following struct has two embedded arrays:
struct ns Contact
{
char *firstName;
char *lastName;
int sizePhones;
ULONG64 *phoneNumber; // array of phone numbers
int sizeEmails;
char **emailAddress; // array of email addresses
char *socSecNumber;
};
The XML serialization of an example ns Contact is:
<mycontact xsi:type="ns:Contact">
<firstName>Joe</firstName>
<lastName>Smith</lastName>
<phoneNumber>5551112222</phoneNumber>
<phoneNumber>5551234567</phoneNumber>
<phoneNumber>5552348901</phoneNumber>
<emailAddress>[email protected]</emailAddress>
<emailAddress>[email protected]</emailAddress>
<socSecNumber>999999999</socSecNumber>
</mycontact>
10.11.8
STL Containers
gSOAP supports the STL containers std::deque, std::list, std::set, and std::vector.
STL containers can only be used within classes to declare members that contain multiple values.
This is somewhat similar to the embedding of arrays in structs in C as explained in Section 10.11.7,
but the STL container approach is more flexible.
You need to import stldeque.h, stllist.h, stlset.h, or stlvector.h to enable std::deque, std::list, std::set, and
std::vector (de)serialization. Here is an example:
#import ”stlvector.h”
class ns myClass
140
{ public:
std::vector<int> *number;
std::vector<xsd string> *name;
...
};
The use of pointer members is not required but advised. The reason is that interoperability with
other SOAP toolkits may lead to copying of ns myClass instances at run time when (de)serializing
multi-referenced data. When a copy is made, certain parts of the containers will be shared between
the copies which could lead to disaster when the classes with their containers are deallocated.
Another way to avoid this is to declare class ns myClass within other data types via a pointer.
(Interoperability between gSOAP clients and services does not lead to copying.)
The XML Schema that corresponds to the ns myClass type is
<complexType name="myClass">
<sequence>
<element name="number" type="xsd:int" minOccurs="1" maxOccurs="unbounded"/>
<element name="name" type="xsd:string" minOccurs="1" maxOccurs="unbounded"/>
...
</sequence>
</complexType>
You can specify the minOccurs and maxOccurs values as explained in Section 18.2.
You can also implement your own containers similar to STL containers. The containers must be
class templates and should define an iterator type, and void clear(), iterator begin(), iterator end(), and
iterator insert(iterator pos, const reference val). The iterator should have a dereference operator to access
the container’s elements. The dereference operator is used by gSOAP to send a sequence of XML
element values. The insert method can be used as a setter method. gSOAP reads a sequence of
XML element values and inserts them in the container via this method.
Here is in example user-defined container template class:
// simpleVector.h
template <class T>
class simpleVector
{
public:
typedef T value type;
typedef value type * pointer;
typedef const value type * const pointer;
typedef value type & reference;
typedef const value type & const reference;
typedef pointer iterator;
typedef const pointer const iterator;
protected:
iterator start;
iterator finish;
size t length;
public:
141
simpleVector() { clear(); }
˜simpleVector() { delete[] start; }
void clear() { start = finish = NULL; }
iterator begin() { return start; }
const iterator begin() const { return start; }
iterator end() { return finish; }
const iterator end() const { return finish; }
size t size() const { return finish-start; }
iterator insert(iterator pos, const reference val)
{
if (!start)
start = finish = new value type[length = 4];
else if (finish >= start + length)
{
iterator i = start;
iterator j = new value type[2 * length];
start = j;
finish = start + length;
length *= 2;
if (pos)
pos = j + (pos - i);
while (i != finish)
*j++ = *i++;
}
if (pos && pos != finish)
{ iterator i = finish;
iterator j = i - 1;
while (j != pos)
*i– = *j–;
}
*finish++ = val;
return pos;
}
};
To enable the container, we add the following two lines to our gSOAP header file:
#include ”simpleVector.h”
template <class T> class simpleVector;
The container class should not be defined in the gSOAP header file. It must be defined in a separate
header file (e.g. ”simpleVector.h”). The template <class T> class simpleVector declaration ensures
that gSOAP will recognize simpleVector as a container class.
Caution: when parsing XML content the container elements may not be stored in the same order
given in the XML content. When gSOAP parses XML it uses the insert container methods to store
elements one by one. However, element content that is “forwarded” with href attributes will be
appended to the container. Forwarding can take place with multi-referenced data that is referred
to from the main part of the SOAP 1.1 XML message to the independent elements that carry
ids. Therefore, your application should not rely on the preservation of the order of elements in a
container.
142
10.11.9
Polymorphic Dynamic Arrays and Lists
Polymorphic arrays (arrays of polymorphic element types) can be encoded when declared as an
array of pointers to class instances. and lists. For example:
class ns Object
{
public:
...
};
class ns Data: public ns Object
{
public:
...
};
class ArrayOfObject
{
public:
ns Object ** ptr; // pointer to array of pointers to Objects
int size; // number of Objects pointed to
int offset; // optional SOAP 1.1 array offset
};
The pointers in the array can point to the ns Object base class or ns Data derived class instances
which will be serialized and deserialized accordingly in SOAP. That is, the array elements are
polymorphic.
10.11.10
How to Change the Tag Names of the Elements of a SOAP Array or List
The ptr field in a struct or class declaration of a dynamic array may have an optional suffix part
that describes the name of the tags of the SOAP array XML elements. The suffix is part of the
field name:
Type * ptrarray elt name
The suffix describes the tag name to be used for all array elements. The usual identifier to XML
translations apply, see Section 9.3. The default XML element tag name for array elements is item
(which corresponds to the use of field name ptritem).
Consider for example:
struct ArrayOfstring
{
xsd string * ptrstring;
int
size; };
The array is serialized as:
<array xsi:type="SOAP-ENC:Array" SOAP-ENC:arrayType="xsd:string[2]">
<string xsi:type="xsd:string">Hello</string>
<string xsi:type="xsd:string">World</string>
</array>
143
SOAP 1.1 and 1.2 do not require the use of a specific tag name for array elements. gSOAP will
deserialize a SOAP array while ignoring the tag names. Certain XML Schemas used in doc/literal
encoding may require the declaration of array element tag names.
10.12
Base64Binary XML Schema Type Encoding
The base64Binary XML Schema type is a special form of dynamic array declared with a pointer
( ptr) to an unsigned char array.
For example using a struct:
struct xsd base64Binary
{
unsigned char * ptr;
int size;
};
Or with a class:
class xsd base64Binary
{
public:
unsigned char * ptr;
int size;
};
When compiled by the gSOAP stub and skeleton compiler, this header file specification will generate
base64Binary serializers and deserializers.
The SOAP ENC:base64 encoding is another type for base 64 binary encoding specified by the SOAP
data type schema and some SOAP applications may use this form (as indicated by their WSDL
descriptions). It is declared by:
struct SOAP ENC base64
{
unsigned char * ptr;
int size;
};
Or with a class:
class SOAP ENC base64
{
unsigned char * ptr;
int size;
};
When compiled by the gSOAP stub and skeleton compiler, this header file specification will generate
SOAP-ENC:base64 serializers and deserializers.
The advantage of using a class is that methods can be used to initialize and manipulate the
and size fields. The user can add methods to this class to do this. For example:
144
ptr
class xsd base64Binary
{
public:
unsigned char * ptr;
int size;
xsd base64Binary(); // Constructor
xsd base64Binary(struct soap *soap, int n); // Constructor
˜xsd base64Binary(); // Destructor
unsigned char *location(); // returns the memory location
int size(); // returns the number of bytes
};
Here are example method implementations:
xsd base64Binary::xsd base64Binary()
{
ptr = NULL;
size = 0;
}
xsd base64Binary::xsd base64Binary(struct soap *soap, int n)
{
ptr = (unsigned char*)soap malloc(soap, n);
size = n;
}
xsd base64Binary::˜xsd base64Binary()
{}
unsigned char *xsd base64Binary::location()
{
return ptr;
}
int xsd base64Binary::size()
{
return size;
}
The following example in C/C++ reads from a raw image file and encodes the image in SOAP
using the base64Binary type:
...
FILE *fd = fopen("image.jpg", "rb");
xsd base64Binary image(&soap, filesize(fd));
fread(image.location(), image.size(), 1, fd);
fclose(fd);
soap begin(&soap);
image.soap serialize(&soap);
image.soap put(&soap, "jpegimage", NULL);
soap end(&soap);
...
where filesize is a function that returns the size of a file given a file descriptor.
Reading the xsd:base64Binary encoded image.
145
...
xsd base64Binary image;
soap begin(&soap);
image.get(&soap, "jpegimage");
soap end(&soap);
...
The struct or class name soap enc base64 should be used for SOAP-ENC:base64 schema type instead
of xsd base64Binary.
10.13
hexBinary XML Schema Type Encoding
The hexBinary XML Schema type is a special form of dynamic array declared with the name
xsd hexBinary and a pointer ( ptr) to an unsigned char array.
For example, using a struct:
struct xsd hexBinary
{
unsigned char * ptr;
int size;
};
Or using a class:
class xsd hexBinary
{
public:
unsigned char * ptr;
int size;
};
When compiled by the gSOAP stub and skeleton compiler, this header file specification will generate
base64Binary serializers and deserializers.
10.14
Literal XML Encoding Style
gSOAP supports document/literal encoding by default. Just as with SOAP RPC encoding, literal
encoding requires the XML Schema of the message data to be provided e.g. in WSDL in order for
the gSOAP compiler to generate the (de)serialization routines. Alternatively, the optional DOM
parser (dom.c and dom++.cpp) can be used to handle generic XML or arbitrary XML documents
can be (de)serialized into regular C strings or wide character strings (wchar t*) by gSOAP (see
Section 10.14.1).
The //gsoap service encoding, //gsoap service method-encoding, and //gsoap service method-response-encoding
directives explicitly enable SOAP encoded or literal encoded messages. For example, to enable RPC
encoding style for the entire service, use:
//gsoap ns service encoding: encoded
146
To enable encoding for particular service methods, use:
//gsoap ns service method-encoding: myMethod encoded
int ns myMethod(...)
To enable encoding for particular service methods responses when the method request is literal,
use:
//gsoap ns service method-response-encoding: myMethod encoded
int ns myMethod(...)
Instead of the encoded value, you can use literal, or a specific encoding style value.
Consider the following example that uses the directive to make the literal encoding explicit. The
LocalTimeByZipCode remote service method of the LocalTime service provides the local time given a
zip code and uses literal encoding (with MS .NET). The following header file declares the method:
int LocalTimeByZipCode(char *ZipCode, char **LocalTimeByZipCodeResult);
Note that none of the data types need to be namespace qualified using namespace prefixes.
//gsoap ns service name: localtime
//gsoap ns service encoding: literal
//gsoap ns service namespace: http://alethea.net/webservices/
int ns LocalTimeByZipCode(char *ZipCode, char **LocalTimeByZipCodeResult);
In this case, the method name requires to be associated with a schema through a namespace prefix,
e.g. ns is used in this example. See Section 18.2 for more details on gSOAP directives. With these
directives, the gSOAP compiler generates client and server sources with the specified settings.
The example client program is:
#include ”soapH.h”
#include ”localtime.nsmap” // include generated map file
int main()
{
struct soap soap;
char *t;
soap init(&soap);
if (soap call ns LocalTimeByZipCode(&soap, ”http://alethea.net/webservices/LocalTime.asmx”,
”http://alethea.net/webservices/LocalTimeByZipCode”, ”32306”, &t))
soap print fault(&soap, stderr);
else
printf(”Time = %s\n”, t);
return 0;
}
To illustrate the manual doc/literal setting, the following client program sets the required properties
before the call:
147
#include ”soapH.h”
#include ”localtime.nsmap” // include generated map file
int main()
{
struct soap soap;
char *t;
soap init(&soap);
soap.encodingStyle = NULL; // don’t use SOAP encoding
soap set omode(&soap, SOAP XML TREE);” // don’t produce multi-ref data (but can accept)
if (soap call ns LocalTimeByZipCode(&soap, ”http://alethea.net/webservices/LocalTime.asmx”,
”http://alethea.net/webservices/LocalTimeByZipCode”, ”32306”, &t))
soap print fault(&soap, stderr);
else
printf(”Time = %s\n”, t);
return 0;
}
The SOAP request is:
POST /webservices/LocalTime.asmx HTTP/1.0
Host: alethea.net
Content-Type: text/xml; charset=utf-8
Content-Length: 479
SOAPAction: "http://alethea.net/webservices/LocalTimeByZipCode"
<?xml version="1.0" encoding="UTF-8"?>
<SOAP-ENV:Envelope
xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:SOAP-ENC="http://schemas.xmlsoap.org/soap/encoding/"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:xsd="http://www.w3.org/2001/XMLSchema"
<SOAP-ENV:Body>
<LocalTimeByZipCode xmlns="http://alethea.net/webservices/">
<ZipCode>32306</ZipCode></LocalTimeByZipCode>
</SOAP-ENV:Body>
</SOAP-ENV:Envelope>
10.14.1
Serializing and Deserializing Mixed Content XML With Strings
To declare a literal XML “type” to hold XML documents in regular strings, use:
typedef char *XML;
To declare a literal XML “type” to hold XML documents in wide character strings, use:
typedef wchar t *XML;
Note: only one of the two storage formats can be used. The differences between the use of regular
strings versus wide character strings for XML documents are:
148
• Regular strings for XML documents MUST hold UTF-8 encoded XML documents. That is,
the string MUST contain the proper UTF-8 encoding to exchange the XML document in
SOAP messages.
• Wide character strings for XML documents SHOULD NOT hold UTF-8 encoded XML documents. Instead, the UTF-8 translation is done automatically by the gSOAP runtime marshalling routines.
Here is a C++ example of a remote method specification in which the parameters of the remote
method uses literal XML encoding to pass an XML document to a service and back:
typedef char *XML;
ns GetDocument(XML m XMLDoc, XML &m XMLDoc );
and in C:
typedef char *XML;
ns GetDocument(XML m XMLDoc, XML *m XMLDoc );
The ns Document is essentially a struct that forms the root of the XML document. The use of the
underscore in the ns Document response part of the message avoids the name clash between the
structs. Assuming that the namespace mapping table contains the binding of ns to http://my.org/
and the binding of m to http://my.org/mydoc.xsd, the XML message is:
<?xml version="1.0" encoding="UTF-8"?>
<SOAP-ENV:Envelope
xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:SOAP-ENC="http://schemas.xmlsoap.org/soap/encoding/"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:xsd="http://www.w3.org/2001/XMLSchema"
xmlns:ns="http://my.org/"
xmlns:m="http://my.org/mydoc.xsd"
SOAP-ENV:encodingStyle="">
<SOAP-ENV:Body>
<ns:GetDocument>
<XMLDoc xmlns="http://my.org/mydoc.xsd">
...
</XMLDoc>
</ns:Document>
</SOAP-ENV:Body>
</SOAP-ENV:Envelope>
When using literal encoding of method parameters and response as shown in the example above,
the literal XML encoding style MUST be specified by setting soap.encodingStyle. For example, to
specify no constraints on the encoding style (which is typical) use NULL:
struct soap soap;
soap init(&soap);
soap.encodingStyle = NULL;
149
As a result, the SOAP-ENV:encodingStyle attribute will not appear in the SOAP payload.
For interoperability with Apache SOAP, use
struct soap soap;
soap init(&soap);
soap.encodingStyle = ”http://xml.apache.org/xml-soap/literalxml”;
When the response parameter is an XML type, it will store the entire XML response content but
without the enveloping response element.
The XML type can be used as part of any data structure to enable the rendering and parsing of
custom XML documents. For example:
typedef char *XML;
struct ns Data /* data in namespace ’ns’ */
{
int number;
char *name;
XML m document; /* XML document in default namespace ’m’ */
};
ns Example(struct ns Data data, struct ns ExampleResponse { struct ns Data data; } *out);
11
SOAP Fault Processing
A predeclared standard SOAP Fault data structure is generated by the gSOAP stub and skeleton
compiler for exchanging exception messages. The built-in struct SOAP ENV Fault data structure is
defined as:
struct SOAP ENV Fault
{
QName faultcode; // QName is builtin
char *faultstring;
char *faultactor;
struct SOAP ENV Detail *detail;
struct SOAP ENV Code *SOAP ENV Code; // MUST be a SOAP ENV Code struct defined
below
char *SOAP ENV Reason;
char *SOAP ENV Node;
char *SOAP ENV Role;
struct SOAP ENV Detail SOAP ENV Detail; // SOAP 1.2 detail field
}; struct SOAP ENV Code
{
QName SOAP ENV Value;
struct SOAP ENV Code *SOAP ENV Subcode; };
struct SOAP ENV Detail
{
int type; // The SOAP TYPE of the object serialized as Fault detail
void *fault; // pointer to the fault object, or NULL
char * any; // any other detail element content (stored in XML format)
};
150
The first four fields in SOAP ENV Fault are SOAP 1.1 specific. The last five fields are SOAP 1.2
specific. You can redefine these structures in the header file. For example, you can use a class for
the SOAP ENV Fault and add methods for convenience.
The data structure content can be changed to the need of an application, but this is generally not
necessary because the application-specific SOAP Fault details can be serialized via the type and
fault fields in the SOAP ENV Detail field, see Section 10.9 on the serialization of data refered to by
type and fault.
The type field allows application data to be serialized as part of the SOAP Fault. The application
data SHOULD be defined as XML elements, which requires you to declare the type names with
a leading underscore to ensure that the types are compatible with XML elements and not just
simpleTypes and complexTypes.
When the skeleton of a remote method returns an error (see Section 9.2), then soap.fault contains
the SOAP Fault data at the receiving side (client).
Server-side faults are raised with soap sender fault or soap receiver fault. The soap sender fault call
should be used to inform that the sender is at fault and the sender (client) should not resend
the request. The soap receiver fault call should be used to indicate a temporary server-side problem,
so a sender (client) can resend the request later. For example:
int ns1 myMethod(struct soap *soap, ...)
{
...
return soap receiver fault(soap, ”Resource temporarily unavailable”, NULL); // return fault to
sender
}
In the example, the SOAP Fault details were empty (NULL). You may pass an XML fragment,
which will be literally included in the SOAP Fault message. For WS-I Basic Profile compliance,
you must pass an XML string with one or more namespace qualified elements, such as:
return soap receiver fault(soap, ”Resource temporarily unavailable”, ”<errorcode xmlns=’http://tempuri.org’>123</err
xmlns=’http://tempuri.org’>abc</errorinfo>”);
When a remote method must raise an exception with application SOAP Fault details, it does so by
assigning the soap.fault field of the current reference to the runtime environment with appropriate
data associated with the exception and by returning the error SOAP FAULT. For example:
soap receiver fault(soap, ”Stack dump”, NULL);
if (soap->version == 2) // SOAP 1.2 is used
{
soap->fault->SOAP ENV Detail = (struct SOAP ENV Detail*)soap malloc(soap, sizeof(struct
SOAP ENV Detail);
soap->fault->SOAP ENV Detail-> type = SOAP TYPE ns1 myStackDataType; // stack
type
soap->fault->SOAP ENV Detail->fault = sp; // point to stack
soap->fault->SOAP ENV Detail-> any = NULL; // no other XML data
}
else
151
{
soap->fault->detail = (struct SOAP ENV Detail*)soap malloc(soap, sizeof(struct SOAP ENV Detail);
soap->fault->detail-> type = SOAP TYPE ns1 myStackDataType; // stack type
soap->fault->detail->fault = sp; // point to stack
soap->fault->detail-> any = NULL; // no other XML data
}
return SOAP FAULT; // return from remote method call
When soap receiver fault allocates a fault struct, this data is removed with the soap end call (or
soap dealloc). Note that the soap receiver fault function is called to allocate the fault struct and set
the fault string and detail fields, i.e. soap receiver fault(soap, ”Stack dump”, NULL). The advantage is
that this is independent of SOAP 1.1 and SOAP 1.2. However, setting the custom detail fields
requires inspecting the SOAP version used, using the soap->version attribute which is 1 for SOAP
1.1 and 2 for SOAP 1.2.
Each remote method implementation in a service application can return a SOAP Fault upon an
exception by returning an error code, see Section 7.2.1 for details and an example. In addition, a
SOAP Fault can be returned by a service application through calling the soap send fault function.
This is useful in case the initialization of the application fails, as illustrated in the example below:
int main()
{
struct soap soap;
soap init(&soap);
some initialization code
if (initialization failed)
{
soap.error = soap receiver fault(&soap, "Init failed", NULL); // set the error condition
(SOAP FAULT)
soap send fault(&soap); // Send SOAP Fault to client
return 0; // Terminate
}
}
12
SOAP Header Processing
A predeclared standard SOAP Header data structure is generated by the gSOAP stub and skeleton
compiler for exchanging SOAP messages with SOAP Headers. This predeclared data structure is:
struct SOAP ENV Header
{ void *dummy;
};
which declares and empty header (some C and C++ compilers don’t accept empty structs so a
transient dummy field is provided).
To adapt the data structure to a specific need for SOAP Header processing, a new struct SOAP ENV Header
can be added to the header file input to the gSOAP compiler. A class for the SOAP Header data
structure can be used instead of a struct.
152
For example, the following header can be used for transaction control:
struct SOAP ENV Header
{ char *t transaction;
};
with client-side code:
struct soap soap;
soap init(&soap);
...
soap.header = NULL; // do not use a SOAP Header for the request (as set with soap init)
soap.actor = NULL; // do not use an actor (receiver is actor)
soap call method(&soap, ...);
if (soap.header) // a SOAP Header was received
cout << soap.header->t transaction;
// Can reset, modify, or set soap.header here before next call
soap call method(&soap, ...); // reuse the SOAP Header of the service response for the request
...
The SOAP Web service response can include a SOAP Header with a transaction number that the
client is supposed to use for the next remote method invocation to the service. Therefore, the next
request includes a transaction number:
...
<SOAP-ENV:Envelope ...>
<SOAP-ENV:Header>
<transaction xmlns="..." xsi:type="int">12345</transaction>
</SOAP-ENV:Header>
<SOAP-ENV:Body>
...
</SOAP-ENV:Body>
</SOAP-ENV:Envelope>
This is just an example and the transaction control is not a feature of SOAP but can be added
on by the application layer to implement stateful transactions between clients and services. At the
client side, the soap.actor attribute can be set to indicate the recipient of the header (the SOAP
SOAP-ENV:actor attribute).
A Web service can read and set the SOAP Header as follows:
int main()
{
struct soap soap;
soap.actor = NULL; // use this to accept all headers (default)
soap.actor = ”http://some/actor”; // accept headers destined for ”http://some/actor” only
soap serve(&soap);
}
...
int method(struct soap *soap, ...)
153
{
if (soap->header) // a Header was received
... = soap->header->t transaction;
else
soap->header = soap malloc(sizeof(struct SOAP ENV Header)); // alloc new header
...
soap->header->t transaction = ...;
return SOAP OK;
}
See Section 18.2 on how to generate WSDL with the proper method-to-header-part bindings.
The SOAP-ENV:mustUnderstand attribute indicates the requirement that the recipient of the SOAP
Header (who must correspond to the SOAP-ENV:actor attribute when present or when the attribute
has the value SOAP-ENV:actor="http://schemas.xmlsoap.org/soap/actor/next") MUST handle the
Header part that carries the attribute. gSOAP handles this automatically on the background.
However, an application still needs to inspect the header part’s value and handle it appropriately.
If a remote method in a Web service is not able to do this, it should return SOAP MUSTUNDERSTAND
to indicate this failure.
The syntax for the header file input to the gSOAP compiler is extended with a special storage
qualifier mustUnderstand. This qualifier can be used in the SOAP Header declaration to indicate
which parts should carry a SOAP-ENV:mustUnderstand=”1” attribute. For example:
struct SOAP ENV Header
{
char *t transaction;
mustUnderstand char *t authentication;
};
When both fields are set and soap.actor=”http://some/actor” then the message contains:
<SOAP-ENV:Envelope ...>
<SOAP-ENV:Header>
<transaction xmlns="...">5</transaction>
<authentication xmlns="..." SOAP-ENV:actor="http://some/actor" SOAP-ENV:mustUnderstand="1">XX
</authentication>
</SOAP-ENV:Header>
<SOAP-ENV:Body>
...
</SOAP-ENV:Body>
</SOAP-ENV:Envelope>
13
MIME Attachments
The gSOAP toolkit supports MIME attachments as per SOAP with Attachments (SwA) specification (http://www.w3.org/TR/SOAP-attachments). In the following discussion, MIME attachment
data is assumed to be resident in memory for sending operations and MIME attachments received
will be stored in memory. MTOM and DIME attachments on the other hand can be streamed and
154
therefore MTOM/DIME attachment data does not need to be stored in memory, see Section 14
and 15.
Transmitting multipart/related MIME attachments with a SOAP/XML message is accomplished
with two functions, soap set mime and soap set mime attachment. The first function is for initialization
purposes and the latter function is used to specify meta data and content data for each attachment.
13.1
Sending a Collection of MIME Attachments (SwA)
The following functions should be used to set up a collection of multipart/related MIME attachments for transmission with a SOAP/XML message.
Function
void soap set mime(struct soap *soap, const char *boundary, const char *start)
This function must be called first to initialize MIME attachment send operations (receives are automatic). The function specifies a MIME boundary and start content ID used for the SOAP message
body. When boundary is NULL, an appropriate MIME boundary will be choosen (important: boundaries cannot occur in the SOAP/XML message and cannot occur in any of the MIME attachments
content). When a specific boundary value is provided, gSOAP will NOT verify that the boundary
is valid. When start is NULL, the start ID of the SOAP message is <SOAP-ENV:Envelope>.
int soap set mime attachment(struct soap *soap, char *ptr, size t size, enum soap mime encoding
encoding, const char *type, const char *id, const char *location, const char *description)
This function adds a new attachment to the list of attachments, where ptr and size refer to the block of memory that holds the attachment data.
The encoding parameter specifies the content encoding of this block, where the value of encoding is one of
SOAP MIME 7BIT, SOAP MIME 8BIT, SOAP MIME BINARY, SOAP MIME QUOTED PRINTABLE,
SOAP MIME BASE64, SOAP MIME IETF TOKEN, or SOAP MIME X TOKEN. These constants reflect the content encoding defined in RFC2045 and you MUST adhere to the content encoding rules
defined by RFC2045. When in doubt, use SOAP MIME BINARY, since this encoding type covers
any content. The mandatory type string parameter is the MIME type of the data. The id string
parameter is the content ID of the MIME attachment. The optional location string parameter is
the content location of the attachment. The optional description string parameter holds a textual
description of the attachment (it may not contain any control characters). All parameter values are
copied, except ptr which must point to a valid location of the attachment data during the transfer.
The value SOAP OK is returned when the attachment was added. Otherwise a gSOAP error code
is returned.
void soap clr mime(struct soap *soap)
Disables MIME attachments, e.g. to avoid MIME attachments to be part of a SOAP Fault response
message.
When providing a MIME boundary with soap set mime, you have to make sure the boundary cannot
match any SOAP/XML message content. Or you can simply pass NULL and let gSOAP select a
safe boundary for you.
The internal list of attachments is destroyed with soap end, you should call this function sometime
after the message exchange was completed (the content of the block of memory referred to by the
ptr parameter is unaffected).
The following example shows how a multipart/related HTTP message with three MIME attachments is set up and transmitted to a server. The first attachment contains the SOAP message.
155
The second and third attachments contain image data. In this example we let the SOAP message
body refer to the attachments using href attributes. The struct claim form data type includes a
definition of a href attribute for this purpose.
struct claim form form1, form2;
form1.href = "cid:[email protected]";
form2.href = "cid:[email protected]";
/* initialize and enable MIME */
soap set mime(soap, "MIME_boundary", "<[email protected]>");
/* add a base64 encoded tiff image (tiffImage points to base64 data) */
soap set mime attachment(soap, tiffImage, tiffLen, SOAP MIME BASE64, "image/tiff", "<claim061400a.tiff@cla
NULL, NULL);
/* add a raw binary jpeg image (jpegImage points to raw data) */
soap set mime attachment(soap, jpegImage, jpegLen, SOAP MIME BINARY, "image/jpeg", "<claim061400a.jpeg@
NULL, NULL);
/* send the forms as MIME attachments with this invocation */
if (soap call claim insurance claim auto(soap, form1, form2, ...))
// an error occurred
else
// process response
Note: the above example assumes that the boundary MIME_boundary does not match any part of
the SOAP/XML message.
The claim form struct is declared in the gSOAP header file as:
struct claim form
{ @char *href;
};
This data type defines the parameter data of the operation. The claim forms in the SOAP/XML
message consist of hrefs to the claim forms attached. The produced message is similar to the
last example shown in the SOAP with Attachments specification (http://www.w3.org/TR/SOAPattachments). Note that the use of href or other attributes for referring to the MIME attachments
is optional according to the SwA standard.
To associate MIME attachments with the request and response of a service operation in the generated WSDL, please see Section 15.1.
The server-side code to transmit MIME attachments back to a client is similar:
int claim insurance claim auto(struct soap *soap, ...)
{
soap set mime(soap, NULL, NULL); // enable MIME
// add a HTML document (htmlDoc points to data, where the HTML doc is stored in compliance
with 7bit encoding RFC2045)
if (soap set mime attachment(soap, htmlDoc, strlen(htmlDoc), SOAP MIME 7BIT, "text/html",
"<[email protected]>", NULL, NULL))
{
soap clr mime(soap); // don’t want fault with attachments
return soap->error;
156
}
return SOAP OK;
}
It is also possible to attach data to a SOAP fault message.
Caution: DIME in MIME is supported. However, gSOAP will not verify whether the MIME
boundary is present in the DIME attachments and therefore will not select a boundary that is
guaranteed to be unique. Therefore, you must provide a MIME boundary with soap set mime that
is unique when using DIME in MIME.
13.2
Retrieving a Collection of MIME Attachments (SwA)
MIME attachments are automatically parsed and stored in memory. After receiving a set of MIME
attachments, either at the client-side or the server-side, the list of MIME attachments can be
traversed to extract meta data and the attachment content. The first attachment in the collection
of MIME attachments always contains meta data about the SOAP message itself (because the
SOAP message was processed the attachment does not contain any useful data).
To traverse the list of MIME attachments in C, you use a loop similar to:
struct soap multipart *attachment;
for (attachment = soap.mime.list; attachment; attachment = attachment->next)
{
printf("MIME attachment:\n");
printf("Memory=%p\n", (*attachment).ptr);
printf("Size=%ul\n", (*attachment).size);
printf("Encoding=%d\n", (int)(*attachment).encoding);
printf("Type=%s\n", (*attachment).type?(*attachment).type:”null”);
printf("ID=%s\n", (*attachment).id?(*attachment).id:”null”);
printf("Location=%s\n", (*attachment).location?(*attachment).location:”null”);
printf("Description=%s\n", (*attachment).description?(*attachment).description:”null”);
}
C++ programmers can use an iterator instead, as in:
for (soap multipart::iterator attachment = soap.mime.begin(); attachment != soap.mime.end();
++attachment)
{
cout << ”MIME attachment:” << endl;
cout << ”Memory=” << (void*)(*attachment).ptr << endl;
cout << ”Size=” << (*attachment).size << endl;
cout << ”Encoding=” << (*attachment).encoding << endl;
cout << ”Type=” << ((*attachment).type?(*attachment).type:”null”) << endl;
cout << ”ID=” << ((*attachment).id?(*attachment).id:”null”) << endl;
cout << ”Location=” << ((*attachment).location?(*attachment).location:”null”) << endl;
cout << ”Description=” << ((*attachment).description?(*attachment).description:”null”) << endl;
}
Note: keep in mind that the first attachment is associated with the SOAP message and you may
want to ignore it.
157
A call to soap end removes all of the received MIME data. To preserve an attachment in memory,
use soap unlink on the ptr field of the soap multipart struct. Recall that the soap unlink function is
commonly used to prevent deallocation of deserialized data.
14
DIME Attachments
The gSOAP toolkit supports DIME attachments as per DIME specification, see http://msdn.microsoft.com/libra
us/dnglobspec/html/draft-nielsen-dime-02.txt
Applications developed with gSOAP can transmit binary DIME attachments with or without
streaming. Without streaming, all data is stored and retrieved in memory, which can be prohibitive when transmitting large files on small devices. In contrast, with DIME streaming, data
handlers are used to pass the data to and from a resource, such as a file or device. With DIME
output streaming, raw binary data is send from a data source in chunks on the fly without buffering
the entire content to save memory. With DIME input streaming, raw binary data will be passed
to data handlers (callbacks).
14.1
Sending a Collection of DIME Attachments
The following functions can be used to explicitly set up a collection of DIME attachments for
transmission with a SOAP/XML message body. The attachments can be streamed, as described
in Section 14.4. Without streaming, each attachment must refer to a block of data in memory.
Function
void soap set dime(struct soap *soap)
This function must be called first to initialize DIME attachment send operations (receives are automatic).
int soap set dime attachment(struct soap *soap, char *ptr, size t size, const char *type, const char
*id, unsigned short optype, const char *option)
This function adds a new attachment to the list of attachments, where ptr and size refer to the block
of memory that holds the attachment data (except when DIME streaming callback handlers are
used as described in Section 14.4. The type string parameter is the MIME type of the data. The
id string parameter is the content ID of the DIME attachment. The option string parameter holds
optional text (gSOAP supports DIME options, but it can send only one) and optype is a user-defined
option type (as per DIME option specification format). All parameter values are copied, except ptr.
The value SOAP OK is returned when the attachment was added. Otherwise a gSOAP error code
is returned.
void soap clr mime(struct soap *soap)
Disables DIME attachments, unless the serialized SOAP message contains attachments for transmission.
These functions allow DIME attachments to be added without requiring message body references.
This is also referred to as the open DIME attachment style. The closed attachment style requires
all DIME attachments to be referenced from the SOAP message body with href (or similar) references. For the closed style, gSOAP supports an automatic binary data serialization method, see
Section 14.3.
158
14.2
Retrieving a Collection of DIME Attachments
DIME attachments are automatically parsed and stored in memory (or passed to the streaming
handlers, when applicable). After receiving a set of DIME attachments, either at the client-side
or the server-side, the list of DIME attachments can be traversed to extract meta data and the
attachment content.
To traverse the list of DIME attachments in C, you use a loop similar to:
struct soap multipart *attachment;
for (attachment = soap.dime.list; attachment; attachment = attachment->next)
{
printf("DIME attachment:\n");
printf("Memory=%p\n", (*attachment).ptr);
printf("Size=%ul\n", (*attachment).size);
printf("Type=%s\n", (*attachment).type?(*attachment).type:”null”);
printf("ID=%s\n", (*attachment).id?(*attachment).id:”null”);
}
C++ programmers can use an iterator instead, as in:
for (soap multipart::iterator attachment = soap.dime.begin(); attachment != soap.dime.end(); ++attachment)
{
cout << ”DIME attachment:” << endl;
cout << ”Memory=” << (void*)(*attachment).ptr << endl;
cout << ”Size=” << (*attachment).size << endl;
cout << ”Type=” << ((*attachment).type?(*attachment).type:”null”) << endl;
cout << ”ID=” << ((*attachment).id?(*attachment).id:”null”) << endl;
}
The options field is available as well. The options content is formatted according to the DIME
specification: the first two bytes are reserved for the option type, the next two bytes store the size
of the option data, followed by the (binary) option data.
A call to soap end removes all of the received DIME data. To preserve an attachment in memory,
use soap unlink on the ptr field of the soap multipart struct. Recall that the soap unlink function is
commonly used to prevent deallocation of deserialized data.
14.3
Serializing Binary Data in DIME
Binary data stored in extended xsd:base64Binary and xsd:hexBinary types can be serialized and
deserialized as DIME attachments. These attachments will be transmitted prior to the sequence of
secondary DIME attachments defined by the user with soap set dime attachment as explained in the
previous section. The serialization process is automated and DIME attachments will be send even
when soap set dime or soap set dime attachment are not used.
The xsd:base64Binary XSD type is defined in gSOAP as a struct or class by
struct xsd base64Binary
{
159
unsigned char * ptr; // pointer to raw binary data
int size; // size of the block of data
};
To enable serialization of the data in DIME, we extend this type with three additional fields:
struct xsd base64Binary
{
unsigned char * ptr;
int size;
char *id;
char *type;
char *options;
};
The three additional fields consist of an id field for attachment referencing (typically a content id
(CID) or UUID), a type field to specify the MIME type of the binary data, and an options field to
piggy-back additional information with a DIME attachment. The order of the declaration of the
fields is significant. In addition, no other fields or methods may be declared before any of these
fields in the struct/class, but additional fields and methods may appear after the field declarations.
An extended xsd hexBinary declaration is similar.
The id and type fields contain text. The set the DIME-specific options field, you can use the
soap dime option function:
char *soap dime option(struct soap *soap, unsigned short type, const char *option)
returns a string with this encoding. For example
struct xsd base64Binary image;
image. ptr = ...;
image. size = ...;
image.id = "uuid:09233523-345b-4351-b623-5dsf35sgs5d6";
image.type = "image/jpeg";
image.options = soap dime option(soap, 0, "My wedding picture");
When either the id or type field values are non-NULL at run time, the data will be serialized as
a DIME attachment. The SOAP/XML message refers to the attachments using href attributes.
This generally works will with SOAP RPC, because href attributes are permitted. However, with
document/literal style the referencing mechanism must be explicitly defined in the schema of the
binary type. The gSOAP declaration of an extended binary type is
struct ns myBinaryDataType
{
unsigned char * ptr;
int size;
char *id;
char *type;
char *options;
};
160
C++ programmers can use inheritance instead of textual extension required in C, as in
class xsd base64Binary
{
unsigned char * ptr;
int size;
};
class ns myBinaryDataType : xsd base64Binary
{
char *id;
char *type;
char *options;
};
This defines an extension of xsd:base64Binary, such that the data can be serialized as DIME
attachments using href attributes for referencing. When a different attribute name is in fact used,
it must be explicitly defined:
//gsoap WSref schema import: http://schemas.xmlsoap.org/ws/2002/04/reference/
struct ns myBinaryDataType
{
unsigned char * ptr;
int size;
char *id;
char *type;
char *options;
@char *WSref location;
};
The example above uses the location attribute defined in the content reference schema, as defined in
one of the vendor’s specific WSDL extensions for DIME (http://www.gotdotnet.com/team/xml wsspecs/dime/W
Extension-for-DIME.htm).
When receiving DIME attachments, the DIME meta data and binary data content is stored in
binary data types only when the XML parts of the message uses href attributes to refer to these
attachments. The gSOAP toolkit may support automatic (de)serialization with other user-defined
(or WSDL-defined) attributes in future releases.
Messages may contain binary data that references external resources not provided as attachments.
In that case, the ptr field is NULL and the id field refers to the external data source.
The dime id format attribute of the current gSOAP run-time environment can be set to the default
format of DIME id fields. The format string MUST contain a %d format specifier (or any other
int-based format specifier). The value of this specifier is a non-negative integer, with zero being the
value of the DIME attachment id for the SOAP message. For example,
struct soap soap;
soap init(&soap);
soap.dime id format = "uuid:09233523-345b-4351-b623-5dsf35sgs5d6-%x";
161
As a result, all attachments with a NULL id field will use a gSOAP-generated id value based on
the format string.
Caution: Care must be taken not to introduce duplicate content id values, when assigning content
id values to the id fields of DIME extended binary data types. Content ids must be unique.
14.4
Streaming DIME
Streaming DIME is achieved with callback functions to fetch and store data during transmission.
Three function callbacks for streaming DIME output and three callbacks for streaming DIME input
are available.
162
Callback (function pointer)
void *(*soap.fdimereadopen)(struct soap *soap, void *handle, const char *id, const char *type, const
char *options)
Called by the gSOAP run-time DIME attachment sender to start reading from a (binary) data
source for outbound transmission. The content will be read from the application’s data source in
chunks using the fdimeread callback and streamed into the SOAP/XML/DIME output stream. The
handle contains the value of the ptr field of an attachment struct/class, which could be a pointer
to specific information such as a file descriptor or a pointer to a string to be passed to this callback.
Both ptr and size fields should have been set by the application prior to the serialization of the
content. The id, type, and options arguments are the DIME id, type, and options, respectively. The
callback should return handle, or another pointer value which will be passed as a handle to fdimeread
and fdimereadclose. The callback should return NULL and set soap->error when an error occurred.
The callback should return NULL (and not set soap->error) when this particular DIME attachment
is not to be streamed.
size t (*soap.fdimeread)(struct soap *soap, void *handle, char *buf, size t len)
Called by the gSOAP run-time DIME attachment sender to read more data from a (binary) data
source for streaming into the output stream. The handle contains the value returned by the fdimereadopen callback. The buf argument is the buffer of length len into which a chunk of data should be
stored. The actual amount of data stored in the buffer may be less than len and this amount
should be returned by the application. A return value of 0 indicates an error (the callback may set
soap->errnum to errno). The size field of the attachment struct/class should have been set by the
application prior to the serialization of the content. The value of size indicates the total size of
the content to be transmitted. When the size is zero then DIME chunked transfers can be used
under certain circumstances to stream content without prior determination of attachment size, see
Section 14.5 below.
void(*soap.fdimereadclose)(struct soap *soap, void *handle)
Called by the gSOAP run-time DIME attachment sender at the end of the streaming process to
close the data source. The handle contains the value returned by the fdimereadopen callback. The
fdimewriteclose callback is called after successfully transmitting the data or when an error occurred.
void *(*soap.fdimewriteopen)(struct soap *soap, const char *id, const char *type, const char *options)
Called by the gSOAP run-time DIME attachment receiver to start writing an inbound DIME attachment to an application’s data store. The content is streamed into an application data store
through multiple fdimewrite calls from the gSOAP attachment receiver. The id, type, and options
arguments are the DIME id, type, and options respectively. The callback should return a handle
which is passed to the fdimewrite and fdimewriteclose callbacks. The ptr field of the attachment
struct/class is set to the value of this handle. The size field is set to the total size of the attachment
after receiving the entire content. The size is unknown in advance because DIME attachments may
be chunked.
int (*soap.fdimewrite)(struct soap *soap, void *handle, const char *buf, size t len)
Called by the gSOAP run-time DIME attachment receiver to write part of an inbound DIME attachment to an application’s data store. The handle contains the value returned by the fdimewriteopen
callback. The buf argument contains the data of length len. The callback should return a gSOAP
error code (e.g. SOAP OK when no error occurred).
void(*soap.fdimewriteclose)(struct soap *soap, void *handle)
Called by the gSOAP run-time DIME attachment receiver at the end of the streaming process to
close the data store. The fdimewriteclose callback is called after successfully receiving the data or
when an error occurred. The handle contains the value returned by the fdimewriteopen callback.
In addition, a void*user field in the struct soap data structure is available to pass user-defined data
to the callbacks. This way, you can set soap.user to point to application data that the callbacks
163
need such as a file name for example.
The following example illustrates the client-side initialization of an image attachment struct to
stream a file into a DIME attachment:
int main()
{
struct soap soap;
struct xsd base64Binary image;
FILE *fd;
struct stat sb;
soap init(&soap);
if (!fstat(fileno(fd), &sb) && sb.st size > 0)
{ // because we can get the length of the file, we can stream it
soap.fdimereadopen = dime read open;
soap.fdimereadclose = dime read close;
soap.fdimeread = dime read;
image. ptr = (unsigned char*)fd; // must set to non-NULL (this is our fd handle which we
need in the callbacks)
image. size = sb.st size; // must set size
}
else
{ // don’t know the size, so buffer it
size t i;
int c;
image. ptr = (unsigned char*)soap malloc(&soap, MAX FILE SIZE);
for (i = 0; i < MAX FILE SIZE; i++)
{
if ((c = fgetc(fd)) == EOF)
break;
image. ptr[i] = c;
}
fclose(fd);
image. size = i;
}
image.type = ”image/jpeg”;
image.options = soap dime option(&soap, 0, ”My picture”);
soap call ns method(&soap, ...);
...
}
void *dime read open(struct soap *soap, void *handle, const char *id, const char *type, const
char *options)
{ return handle;
}
void dime read close(struct soap *soap, void *handle)
{ fclose((FILE*)handle);
}
size t dime read(struct soap *soap, void *handle, char *buf, size t len)
{ return fread(buf, 1, len, (FILE*)handle);
}
The following example illustrates the streaming of a DIME attachment into a file by a client:
164
int main()
{ struct soap soap;
soap init(&soap);
soap.fdimewriteopen = dime write open;
soap.fdimewriteclose = dime write close;
soap.fdimewrite = dime write;
soap call ns method(&soap, ...);
...
}
void *dime write open(struct soap *soap, const char *id, const char *type, const char *options)
{
FILE *handle = fopen(”somefile”, ”wb”);
if (!handle)
{
soap->error = SOAP EOF;
soap->errnum = errno; // get reason
}
return (void*)handle;
}
void dime write close(struct soap *soap, void *handle)
{ fclose((FILE*)handle);
}
int dime write(struct soap *soap, void *handle, const char *buf, size t len)
{
size t nwritten;
while (len)
{
nwritten = fwrite(buf, 1, len, (FILE*)handle);
if (!nwritten)
{
soap->errnum = errno; // get reason
return SOAP EOF;
}
len -= nwritten;
buf += nwritten;
}
return SOAP OK;
}
Note that compression can be used with DIME to compress the entire message. However, compression requires buffering to determine the HTTP content length header, which cancels the
benefits of streaming DIME. To avoid this, you should use chunked HTTP (with the outputmode SOAP IO CHUNK flag) with compression and streaming DIME. At the server side, when
you set SOAP IO CHUNK before calling soap serve, gSOAP will automatically revert to buffering
(SOAP IO STORE flag is set). You can check this flag with (soap-¿omode & SOAP IO) == SOAP IO CHUNK
to see if the client accepts chunking. More information about streaming chunked DIME can be
found in Section 14.5.
Caution: The options field is a DIME-specific data structure, consisting of a 4 byte header containing
the option type info (hi byte, lo byte), option string length (hi byte, lo byte), followed by a non-’\0’
terminated string. The gSOAP DIME handler recognizes one option at most.
165
14.5
Streaming Chunked DIME
gSOAP automatically handles inbound chunked DIME attachments (streaming or non-streaming).
To transmit outbound DIME attachments, the attachment sizes MUST be determined in advance
to calculate HTTP message length required to stream DIME over HTTP. However, gSOAP also
supports the transmission of outbound chunked DIME attachments without prior determination
of DIME attachment sizes when certain conditions are met. These conditions require either nonHTTP transport (use the output-mode SOAP ENC XML flag), or chunked HTTP transport (use the
output-mode SOAP IO CHUNK flag). You can also use the SOAP IO STORE flag (which is also used
automatically with compression to determine the HTTP content length header) but that cancels
the benefits of streaming DIME.
To stream chunked DIME, set the size field of an attachment to zero and enable HTTP chunking.
The DIME fdimeread callback then fetches data in chunks and it is important to fill the entire buffer
unless the end of the data has been reached and the last chunk is to be send. That is, fdimeread
should return the value of the last len parameter and fill the entire buffer buf for all chunks except
the last.
14.6
WSDL Bindings for DIME Attachments
The wsdl2h WSDL parser recognizes DIME attachments and produces an annotated header file.
Both open and closed layouts are supported for transmitting DIME attachments. For closed formats, all DIME attachments must be referenced from the SOAP message, e.g. using hrefs with
SOAP encoding and using the application-specific reference attribute included in the base64Binary
struct/class for doc/lit.
As of this writing, the gSOAP compiler soapcpp2 does not yet produce a WSDL with DIME extensions.
15
MTOM Attachments
MTOM (Message Transmission Optimization Mechanism) is a relatively new format for transmitting attachments with SOAP messages (see http://www.w3.org/TR/soap12-mtom). MTOM is a
W3C working draft as of this writing. MTOM attachments are essentially MIME attachments with
standardized mechanisms for cross referencing attachments from the SOAP body, which is absent
in (plain) MIME attachments and optional with DIME attachments.
Unlike the name suggests, the speed by which attached data is transmitted is not increased compared to MIME, DIME, or even XML encoded base64 data (at least the performance differences
in gSOAP will be small). The advantage of the format is the standardized attachment reference
mechanism, which should improve interoperability.
The MTOM specification mandates SOAP 1.2 and the use of the XOP namespace. The XOP
Include element xop:Include is used to reference attachment(s) from the SOAP message body.
Because references from within the SOAP message body to attachments are mandatory with
MTOM, the implementation of the serialization and deserialization of MTOM MIME attachments
166
in gSOAP uses the extended binary type comparable to DIME support in gSOAP. This binary type
is predefined in the import/xop.h file:
//gsoap xop schema import: http://www.w3.org/2004/08/xop/include
struct xop Include
{
unsigned char * ptr;
int size;
char *id;
char *type;
char *options;
};
typedef struct xop Include xop Include;
The additional id, type, and option fields enable MTOM attachments for the data pointed to by ptr
of size size. The process for sending and receiving MTOM XOP attachments is fully automated.
Streaming techniques however, such as for DIME are not (yet) available. The id field references the
attachment (typically a content id CID or UUID). When set to NULL, gSOAP assigns a unique
CID. The type field specifies the required MIME type of the binary data, and the optional options
field can be used to piggy-back descriptive text with an attachment. The order of the declaration
of the fields is significant.
You can explicitly import the xop.h in your header file to use the MTOM attachments in your
service, for example:
#import ”import/soap12.h”
/* alternatively, without the import above, use:
//gsoap SOAP-ENV schema namespace: http://www.w3.org/2003/05/soap-envelope
//gsoap SOAP-ENC schema namespace: http://www.w3.org/2003/05/soap-encoding
/
#import ”import/xop.h”
#import ”import/xmlmime.h”
//gsoap x schema namespace: http://my.first.mtom.net
struct x myData
{
xop Include xop Include; // attachment
@char *xmlmime contentType; // and its contentType
};
int x myMTOMtest(struct x myData *in, struct x myData *out);
As you can see, there is really no difference between the specification of MTOM and DIME attachments in a gSOAP header file. Except that you MUST use SOAP 1.2 and the xop Include
element.
When an instance of x myDataType is serialized and either or both the id and type fields are nonNULL, the data is transmitted as MTOM MIME attachment if the SOAP ENC MTOM flag is set in
the gSOAP’s soap struct context:
struct soap *soap = soap new1(SOAP ENC MTOM);
167
Without this flag, the attachments will be transmitted in DIME format (Section 14). If your current
clients and services are based on non-streaming DIME attachments using the SOAP body reference
mechanism (thus, without using the soap set dime attachment function) or plain base64 binary XML
data elements, it is very easy to adopt MTOM by renaming the binary types to xop Include and
using the SOAP ENC MTOM flag with the SOAP 1.2 namespace.
15.1
Generating MultipartRelated MIME Attachment Bindings in WSDL
To generate multipartRelated bindings in the WSDL file, use the //gsoap ... service method-mime-type
directive (see also Section 18.2. The directive can be repeated for each attachment you want to
associate with a method’s request and response messages.
For example:
#import ”import/soap12.h”
#import ”import/xop.h”
#import ”import/xmlmime.h”
//gsoap x schema namespace: http://my.first.mtom.net
struct x myData
{
xop Include xop Include; // attachment
@char *xmlmime contentType; // and its contentType
};
//gsoap x service method-mime-type: myMTOMtest text/xml
int x myMTOMtest(struct x myData *in, struct x myData *out);
The //gsoap x service method-mime-type directive indicates that this operation accepts text/xml MIME
attachments. See the SOAP-with-Attachment specification for the MIME types to use (for example,
*/* is a wildcard). If the operation has more than one attachment, just repeat this directive for
each attachment you want to bind to the operation.
To bind attachments only to the request message of an operation, use //gsoap x service method-inputmime-type. Similarly, to bind attachments only to the response message of an operation, use //gsoap
x service method-ouput-mime-type.
The wsdl2h WSDL parser recognizes MIME attachments and produces an annotated header file.
However, the ordering of MIME parts in the multipartRelated elements is not reflected in the
header file. Application developers should adhere the standards and ensure that multipart/related
attachments are transmitted in compliance with the WSDL operation declarations.
15.2
Sending and Receiving MTOM Attachments
A receiver must be informed to recognize MTOM attachments by setting the SOAP ENC MTOM flag
of the gSOAP context. Otherwise, the regular MIME attachment mechanism (SwA) will be used
to store attachments.
When using wsdl2h to build clients and/or services, you should use the typemap.dat file included
in the distribution package. The typemap.dat file defines the XOP namespace and XML MIME
168
namespaces as imported namespaces:
xop = <http://www.w3.org/2004/08/xop/include>
xmime = <http://www.w3.org/2004/06/xmlmime>
xmlmime = <http://www.w3.org/2004/11/xmlmime>
The wsdl2h tool uses the typemap.dat file (see also option -t) to convert WSDL into a gSOAP header
file. In this case we don’t want the wsdl2h tool to read the XOP schema and translate it, since we
have a pre-defined xop Include element to handle XOP for MTOM. This xop Include element is
defined in xop.h. Therefore, the bindings shown above will not translate the XOP and XML MIME
schemas to code, but generates #import statements instead:
#import ”xop.h”
#import ”xmlmime.h”
The #import statements are only added for those namespaces that are actually used by the service.
Let’s take a look at an example. The wsdl2h importer generates a header file with #import ”xop.h”
from a WSDL that references XOP, for example:
#import ”xop.h”
#import ”xmlmime.h”
struct ns Data
{
xop Include xop Include;
@char *xmlmime contentType;
};
Suppose the WSDL defines an operation:
int ns echoData(struct ns Data *in, struct ns Data *out);
After generating the stubs/proxies with the soapcpp2 compiler, we can invoke the stub at the client
side with:
struct soap *soap = soap new1(SOAP ENC MTOM);
struct ns Data data;
data.xop Include. ptr = (unsigned char*)”<b>Hello world!</b>”;
data.xop Include. size = 20;
data.xop Include.id = NULL; // CID automatically generated by gSOAP engine
data.xop Include.type = ”text/html”; // MIME type
data.xop Include.options = NULL; // no descriptive info added
data.xmlmime contentType = ”text/html”; // MIME type
if (soap call ns echoData(soap, endpoint, action, &data, &data)) soap print fault(soap, stderr);
else
printf(”Got data\n”);
soap destroy(soap); // remove deserialized class instances
soap end(soap); // remove temporary and deserialized data
soap free(soap); // detach and free context
169
Note that the xop Include.type field must be set to transmit MTOM attachments, otherwise plain
base64 XML will be used.
At the server side, we show an example of an operation handler that just copies the input data to
output:
int ns echoData(struct soap *soap, struct ns Data *in, struct ns data *out)
{
*out = *in;
return SOAP OK;
}
The server must use the SOAP ENC MTOM flag to initialize the soap struct to receive and send
MTOM attachments.
15.3
Streaming MTOM/MIME
Streaming MTOM/MIME is achieved with callback functions to fetch and store data during transmission. Three function callbacks for streaming MTOM/MIME output and three callbacks for
streaming MTOM/MIME input are available.
170
Callback (function pointer)
void *(*soap.fmimereadopen)(struct soap *soap, void *handle, const char *id, const char *type, const
char *description)
Called by the gSOAP run-time MTOM/MIME attachment sender to start reading from a (binary)
data source for outbound transmission. The content will be read from the application’s data source
in chunks using the fmimeread callback and streamed into the SOAP/XML/MTOM/MIME output
stream. The handle contains the value of the ptr field of an attachment struct/class, which could
be a pointer to specific information such as a file descriptor or a pointer to a string to be passed
to this callback. Both ptr and size fields should have been set by the application prior to the
serialization of the content. The id, type, and description arguments are the MTOM/MIME id, type,
and description, respectively. The callback should return handle, or another pointer value which will
be passed as a handle to fmimeread and fmimereadclose. The callback should return NULL and set
soap->error when an error occurred. The callback should return NULL (and not set soap->error)
when this particular MTOM/MIME attachment is not to be streamed.
size t (*soap.fmimeread)(struct soap *soap, void *handle, char *buf, size t len)
Called by the gSOAP run-time MTOM/MIME attachment sender to read more data from a (binary)
data source for streaming into the output stream. The handle contains the value returned by the
fmimereadopen callback. The buf argument is the buffer of length len into which a chunk of data
should be stored. The actual amount of data stored in the buffer may be less than len and this
amount should be returned by the application. A return value of 0 indicates an error (the callback
may set soap->errnum to errno). The size field of the attachment struct/class should have been
set by the application prior to the serialization of the content. The value of size indicates the
total size of the content to be transmitted. When the size is zero then MTOM/MIME chunked
transfers can be used under certain circumstances to stream content without prior determination of
attachment size, see Section 15.5 below.
void(*soap.fmimereadclose)(struct soap *soap, void *handle)
Called by the gSOAP run-time MTOM/MIME attachment sender at the end of the streaming process
to close the data source. The handle contains the value returned by the fmimereadopen callback. The
fmimewriteclose callback is called after successfully transmitting the data or when an error occurred.
void *(*soap.fmimewriteopen)(struct soap *soap, void *handle, const char *id, const char *type,
const char *description, enum soap mime encoding encoding)
Called by the gSOAP run-time MTOM/MIME attachment receiver to start writing an inbound MTOM/MIME attachment to an application’s data store. The content is streamed
into an application data store through multiple fmimewrite calls from the gSOAP attachment
receiver. The handle argument is normally NULL, unless soap get mime attachment is used
that passes the handle to the callback, see Section 15.4. The id, type, and description arguments are the MTOM/MIME id, type, and description respectively. The encoding enumeration value indicates the MIME content transfer encoding, which is one of SOAP MIME NONE,
SOAP MIME 7BIT, SOAP MIME 8BIT, SOAP MIME BINARY, SOAP MIME QUOTED PRINTABLE,
SOAP MIME BASE64, SOAP MIME IETF TOKEN, SOAP MIME X TOKEN. Content decoding may
have to be considered by the application based on this value. The callback should return a nonNULL handle which is passed to the fmimewrite and fmimewriteclose callbacks. The ptr field of
the attachment struct/class is set to the value of this handle. The size field is set to the total
size of the attachment after receiving the entire content. The size is unknown in advance because
MTOM/MIME attachments may be chunked.
int (*soap.fmimewrite)(struct soap *soap, void *handle, const char *buf, size t len)
Called by the gSOAP run-time MTOM/MIME attachment receiver to write part of an inbound
MTOM/MIME attachment to an application’s data store. The handle contains the value returned
by the fmimewriteopen callback. The buf argument contains the data of length len. The callback
should return a gSOAP error code (e.g. SOAP OK when no error occurred).
void(*soap.fmimewriteclose)(struct soap *soap, void *handle)
Called by the gSOAP run-time MTOM/MIME attachment receiver at the end of the streaming
process to close the data store. The fmimewriteclose callback is called after successfully receiving
171contains the value returned by the fmimewriteopen
the data or when an error occurred. The handle
callback.
In addition, a void*user field in the struct soap data structure is available to pass user-defined data
to the callbacks. This way, you can set soap.user to point to application data that the callbacks
need such as a file name for example.
The following example illustrates the client-side initialization of an image attachment struct to
stream a file into a MTOM attachment without HTTP chunking (HTTP streaming chunked MTOM
transfer is presented in Section 15.5):
int main()
{
struct soap soap;
struct xsd base64Binary image;
FILE *fd;
struct stat sb;
soap init1(&soap, SOAP ENC MTOM); // mandatory to enable MTOM
if (!fstat(fileno(fd), &sb) && sb.st size > 0)
{ // because we can get the length of the file, we can stream it without chunking
soap.fmimereadopen = mime read open;
soap.fmimereadclose = mime read close;
soap.fmimeread = mime read;
image. ptr = (unsigned char*)fd; // must set to non-NULL (this is our fd handle which we
need in the callbacks)
image. size = sb.st size; // must set size
}
else
{ // don’t know the size, so buffer it
size t i;
int c;
image. ptr = (unsigned char*)soap malloc(&soap, MAX FILE SIZE);
for (i = 0; i < MAX FILE SIZE; i++)
{
if ((c = fgetc(fd)) == EOF)
break;
image. ptr[i] = c;
}
fclose(fd);
image. size = i;
}
image.type = ”image/jpeg”; // MIME type
image.options = ”This is my picture”; // description of object
soap call ns method(&soap, ...);
...
}
void *mime read open(struct soap *soap, void *handle, const char *id, const char *type, const
char *description)
{ return handle;
}
void mime read close(struct soap *soap, void *handle)
{ fclose((FILE*)handle);
}
size t mime read(struct soap *soap, void *handle, char *buf, size t len)
{ return fread(buf, 1, len, (FILE*)handle);
172
}
The following example illustrates the streaming of a MTOM/MIME attachment into a file by a
client:
int main()
{ struct soap soap;
soap init(&soap);
soap.fmimewriteopen = mime write open;
soap.fmimewriteclose = mime write close;
soap.fmimewrite = mime write;
soap call ns method(&soap, ...);
...
}
void *mime write open(struct soap *soap, const char *id, const char *type, const char *description, enum soap mime encoding encoding)
{
FILE *handle = fopen(”somefile”, ”wb”);
// We ignore the MIME content transfer encoding here, but should check
if (!handle)
{
soap->error = SOAP EOF;
soap->errnum = errno; // get reason
}
return (void*)handle;
}
void mime write close(struct soap *soap, void *handle)
{ fclose((FILE*)handle);
}
int mime write(struct soap *soap, void *handle, const char *buf, size t len)
{
size t nwritten;
while (len)
{
nwritten = fwrite(buf, 1, len, (FILE*)handle);
if (!nwritten)
{
soap->errnum = errno; // get reason
return SOAP EOF;
}
len -= nwritten;
buf += nwritten;
}
return SOAP OK;
}
Note that compression can be used with MTOM/MIME to compress the entire message. However,
compression requires buffering to determine the HTTP content length header, which cancels the
benefits of streaming MTOM/MIME. To avoid this, you should use chunked HTTP (with the
output-mode SOAP IO CHUNK flag) with compression and streaming MTOM/MIME. At the server
side, when you set SOAP IO CHUNK before calling soap serve, gSOAP will automatically revert to
173
buffering (SOAP IO STORE flag is set). You can check this flag with (soap-¿omode & SOAP IO) ==
SOAP IO CHUNK to see if the client accepts chunking. More information about streaming chunked
MTOM/MIME can be found in Section 15.5.
15.4
Redirecting Inbound MTOM/MIME Streams Based on SOAP Body Content
When it is preferable or required to redirect inbound MTOM/MIME attachment streams based on
SOAP message body content, where for example the names of the resources are listed in the SOAP
message body, an alternative mechanism must be used. This mechanism can be used both at the
client and server side.
Because the routing of the streams is accomplished with explicit function calls, this method should
only be used when required and should not be considered optional. That is, when you enable this
method, you MUST check for pending MTOM/MIME attachments and handle them appropriately.
This is true even when you don’t expect MTOM/MIME attachments in the payload, because the
peer may trick you by sending attachments anyway and you should be prepared to accept or reject
them.
The explicit MTOM/MIME streaming mechanism consists of three API functions:
Function
void soap post check mime attachments(struct soap *soap)
Enables post-message body inbound streaming MTOM/MIME attachments. The presence of attachments must be explicitly checked using the function below.
int soap check mime attachments(struct soap *soap)
Should be called after a client-side call (e.g. soap call ns method) to check the presence of attachments. Returns 1 (true) when attachments are present. If present, each attachment MUST be
processed with the function below.
struct soap multipart *soap get mime attachment(struct soap *soap, void *handle)
Parses an attachment and invokes the MIME callbacks (when set). The handle parameter is passed
to fmimewriteopen. The handle may contain any data that is extracted from the SOAP message
body to guide the redirection of the stream in the callbacks. The return value is a struct with a
char *ptr field that contains the handle value returned by the fmimewriteopen callback, and char *id,
char *type, and char *description fields with the optional MIME id, type, and description info.
Example client-side code in C:
struct soap *soap = soap new1(SOAP ENC MTOM);
soap post check mime attachments(soap);
...
if (soap call ns myMethod(soap, ...))
soap print fault(soap, stderr); // an error occurred
else
{
if (soap check mime attachments(soap)) { // attachments are present, channel is still open
{
do
{
174
... // get data ’handle’ from SOAP response and pass to callbacks
... // set the fmime callbacks, if needed
struct soap multipart *content = soap get mime attachment(soap, (void*)handle);
printf("Received attachment with id=%s and type=%s\n", content->id?content->id:””,
content->type?content->type:””);
} while (content);
if (soap->error)
soap print fault(soap, stderr);
}
}
}
...
soap destroy(soap);
soap end(soap);
soap free(soap); // detach and free context
The server-side service operations are implemented as usual, but with additional checks for MTOM/MIME
attachments:
struct soap *soap = soap new1(SOAP ENC MTOM);
soap post check mime attachments(soap);
...
soap serve(soap);
...
int ns myMethod(struct soap *soap, ...)
{ ... // server-side processing logic
if (soap check mime attachments(soap)) { // attachments are present, channel is still open
{
do
{
... // get data ’handle’ from SOAP request and pass to callbacks
... // set the fmime callbacks, if needed
struct soap multipart *content = soap get mime attachment(soap, (void*)handle);
printf("Received attachment with id=%s and type=%s\n", content->id?content->id:””,
content->type?content->type:””);
} while (content);
if (soap->error)
return soap->error;
}
}
... // server-side processing logic
return SOAP OK;
}
15.5
Streaming Chunked MTOM/MIME
gSOAP automatically handles inbound chunked MTOM/MIME attachments (streaming or nonstreaming). To transmit outbound MTOM/MIME attachments, the attachment sizes MUST be
determined in advance to calculate HTTP message length required to stream MTOM/MIME over
175
HTTP. However, gSOAP also supports the transmission of outbound chunked MTOM/MIME attachments without prior determination of MTOM/MIME attachment sizes when certain conditions
are met. These conditions require either non-HTTP transport (use the output-mode SOAP ENC XML
flag), or chunked HTTP transport (use the output-mode SOAP IO CHUNK flag). You can also use
the SOAP IO STORE flag (which is also used automatically with compression to determine the HTTP
content length header) but that cancels the benefits of streaming MTOM/MIME.
To stream chunked MTOM/MIME, set the size field of an attachment to zero and enable HTTP
chunking. The MTOM/MIME fmimeread callback then fetches data in chunks of any size between
1 and the value of the len argument. The fmimeread callback should return 0 upon reaching the end
of the data stream.
16
XML Validation
The gSOAP XML parser applies basic rules to validate content. However, occurrence constraints are
not automatically verified unless explicitly indicated. This helps to avoid interoperability problems
with toolkits that do not strictly enforce validation rules. In addition, we cannot always use
strict validation for SOAP RPC encoded messages, since SOAP RPC encoding adopts a very loose
serialization format.
Validation constraints are checked by gSOAP with the SOAP XML STRICT input mode flag set,
e.g. with soap set imode(soap, SOAP XML STRICT) or soap new(SOAP XML STRICT), see Section 8.12
for a complete list of flags.
16.1
Occurrence Constraints
16.1.1
Elements with minOccurs and maxOccurs Restrictions
To force the validation of minOccurs and maxOccurs contraints the SOAP XML STRICT input mode
flag must be set. The minOccurs and maxOccurs constraints are specified for fields of a struct and
members of a class in a header file using the following syntax:
Type fieldname [minOccurs[:maxOccurs]] [= value]
The minOccurs and maxOccurs values must be integer literals. A default value can be provided
when minOccurs is zero. Default values must be primitive types, integer, float, string, etc. By
default the minOccurs constraint is zero.
For example
struct ns MyRecord
{
int n = 5; // element with default value 5, minOccurs=0, maxOccurs=1
int m 1; // element with minOccurs=1
int size 0:10; // sequence ¡item¿ with minOccurs=0, maxOccurs=10
int *item;
std::vector<double> nums 2; // sequence ¡nums¿ with minOccurs=2, maxOccurs=unbounded
176
};
struct arrayOfint
{
int * ptr 1:100; // minOccurs=1, maxOccurs=100
int size;
};
Pointer-based struct fields and class members are allowed to be nillable when minOccurs is zero.
16.1.2
Required and Prohibited Attributes
Similar to the minOccurs and maxOccurs annotations defined in the previous section, attributes in
a struct or class can be annotated with occurrence constraints to make them optional (0), required
(1), or prohibited (0:0). Default values can be assigned to optional attributes.
For example
struct ns MyRecord
{
@int m 1; // required attribute (occurs at least once)
@int n = 5; // optional attribute with default value 5
@int o 0; // optional attribute (may or may not occur)
@int p 0:0; // prohibited attribute
};
Remember to set the SOAP XML STRICT input mode flag to enable the validation of attribute
occurrence constraints.
16.1.3
Data Length Restrictions
A schema simpleType is defined with a typedef by taking a base primitive to defined a derived
simpleType. For example:
typedef int time seconds;
This defines the following schema type in time.xsd:
<simpleType name="seconds">
<restriction base="xsd:int"/>
</simpleType>
A complexType with simpleContent is defined with a wrapper struct/class:
struct time date
{
char * item; // some custom format date (restriction of string)
@enum time zone { EST, GMT, ... } zone;
}
177
This defines the following schema type in time.xsd:
<complexType name="date">
<simpleContent>
<extension base="xsd:string"/>
</simpleContent>
<attribute name="zone" type="time:zone" use="optional"/>
</complexType> <simpleType name="zone">
<restriction base="xsd:string">
<enumeration value="EST"/>
<enumeration value="GMT"/>
...
</restriction>
</simpleType>
Data value length constraints of simpleTypes and complexTypes with simpleContent are defined
as follows.
typedef char *ns string256 0:256; // simpleType restriction of string with max length 256 characters
typedef char *ns string10 10:10; // simpleType restriction of string with length of 10 characters
typedef std::string *ns string8 8; // simpleType restriction of string with at least 8 characters
struct ns data
{
char * item :256; // simpleContent with at most 256 characters
@char *name 1; // required name attribute
}; struct time date
{
char * item :100; @enum time zone { EST, GMT, ... } zone = GMT;
}
Remember to set the SOAP XML STRICT input mode flag to enable the validation of value length
constraints.
16.2
Other Constraints
To associate a pattern with a simpleType, you can define a simpleType with a typedef and a pattern
string:
typedef int time second ”[1-5]?[0-9]—60”;
This defines the following schema type in time.xsd:
<simpleType name="second">
<restriction base="xsd:int">
<pattern value="[1-5]?[0-9]|60"/>
</restriction base="xsd:int"/>
</simpleType>
The pattern string MUST contain a valid regular expression.
Patterns are currently not checked in the validation process.
178
17
SOAP-over-UDP
UDP is a simple, unreliable datagram protocol: UDP sockets are connectionless. UDP address
formats are identical to those used by TCP. In particular UDP provides a port identifier in addition
to the normal Internet address format. The UDP port space is separate from the TCP port space
(i.e. a UDP port may not be “connected” to a TCP port). In addition broadcast packets may be
sent (assuming the underlying network supports this) by using a reserved “broadcast address”; this
address is network interface dependent.
To enable SOAP-over-UDP with gSOAP, stdsoap2.c and stdsoap2.cpp MUST be compiled with DWITH UDP prior to linking with your application.
Client-side messages with SOAP-over-UDP endpoint URLs (soap.udp://...) will be automatically
transmitted as datagrams. Server-side applications should set the SOAP IO UDP mode flag to accept
UDP requests, e.g. using soap init1 or soap set mode.
The maximum message length for datagram packets is restricted by the buffer size SOAP BUFLEN,
which is 65536 by default, unless compiled with WITH LEAN to support small-scale embedded systems. For UDP transport SOAP BUFLEN must not exceed the maximum UDP packet size 65536
(the size of datagram messages is constrained by the UDP packet size 216 = 65536 as per UDP
standard). You can use gzip compression to reduce the message size, but note that compressed
SOAP-over-UDP is a gSOAP-specific feature because it is not part of the SOAP-over-UDP specification.
The SOAP-over-UDP specification relies on WS-Addressing. The wsa.h file in the import directory
defines the WS-Addressing elements for client and server applications.
The gSOAP implementation conforms to the SOAP-over-UDP requirements:
• SOAP-over-UDP server endpoint URL format: soap.udp://host:port/path
• Support one-way message-exchange pattern (MEP) where a SOAP envelope is carried in a
user datagram.
• Support request-response message-exchange pattern (MEP) where SOAP envelopes are carried in user datagrams.
• Support multicast transmission of SOAP envelopes carried in user datagrams.
• Support both SOAP 1.1 and SOAP 1.2 envelopes.
The following additional features are also available, but are not supported by the SOAP-over-UDP
specification:
• Zlib/gzip message compression (compile -DWITH GZIP).
• SOAP with DIME attachments over UDP.
• SOAP with MIME attachments (SwA) over UDP.
• Support for IPv6 (compile -DWITH IPV6)
179
17.1
Using WS-Addressing with SOAP-over-UDP
A SOAP-over-UDP application MUST use WS-Addressing to control message delivery as per
SOAP-over-UDP specification.
The wsa.h file in the import directory defines the WS-Addressing elements. To include the WSAddressing elements in the SOAP Header for messaging, a struct SOAP ENV Header structure must
be defined in your header file with the appropriate WS-Addressing elements. For example:
#import ”wsa.h”
struct SOAP ENV
{
mustUnderstand
mustUnderstand
mustUnderstand
mustUnderstand
mustUnderstand
mustUnderstand
mustUnderstand
};
Header
wsa
wsa
wsa
wsa
wsa
wsa
wsa
MessageID wsa MessageID 0;
RelatesTo *wsa RelatesTo 0;
From *wsa From 0;
ReplyTo *wsa ReplyTo 0;
FaultTo *wsa FaultTo 0;
To wsa To 0;
Action wsa Action 0;
We also included a //gsoap wsa schema import directive in the wsa.h file to enable the generation of
WSDL specifications that import (instead of includes) the WS-Addressing elements. Note that the
//gsoapopt w directive must not be present in your header file to enable WSDL generation.
One-way SOAP-over-UDP messages (see Section 7.3) should be declared to include the wsa:MessageID,
wsa:To, and wsa:Action elements in the SOAP Header of the request message as follows:
//gsoap ns service method-header-part: sendString wsa MessageID
//gsoap ns service method-header-part: sendString wsa To
//gsoap ns service method-header-part: sendString wsa Action
int ns sendString(char *str, void);
Request-response SOAP-over-UDP messages should be declared to include the wsa:MessageID, wsa:To,
wsa:Action, and wsa:ReplyTo elements in the SOAP Header of the request message, and the the
wsa:MessageID, wsa:To, wsa:Action, and wsa:RelatesTo elements in the SOAP Header of the response
message:
//gsoap ns service method-header-part: echoString wsa MessageID
//gsoap ns service method-header-part: echoString wsa To
//gsoap ns service method-header-part: echoString wsa Action
//gsoap ns service method-input-header-part: sendString wsa ReplyTo
//gsoap ns service method-output-header-part: echoString wsa RelatesTo
int ns echoString(char *str, char **res);
For the content requirements of these elements, please consult the SOAP-over-UDP specification
and/or read the next sections explaining SOAP-over-UDP unicast, multicast, one-way, and requestresponse client and server applications.
180
17.2
Client-side One-way Unicast
This example assumes that the gSOAP header file includes the SOAP Header with WS-Addressing
elements and the ns sendString function discussed in Section 17.1
struct soap soap;
struct SOAP ENV Header header; // the SOAP Header
soap init(&soap);
soap.send timeout = 1; // 1s timeout
soap default SOAP ENV Header(&soap, &header); // init SOAP Header
header.wsa MessageID = ”message ID”;
header.wsa To = ”server URL”;
header.wsa Action = ”server action”;
soap.header = &header; // bind the SOAP Header for transport
// Send the message over UDP:
if (soap send ns echoString(&soap, ”soap.udp://...”, NULL, ”hello world!”))
soap print fault(&soap, stderr); // report error
soap end(&soap); // cleanup
soap destroy(&soap); // cleanup
soap done(&soap); // close connection (should not use soap struct after this)
17.3
Client-side One-way Multicast
This example is similar to the one-way unicast example discussed above, but uses a broadcast
address and the SO BROADCAST socket option:
struct soap soap;
struct SOAP ENV Header header; // the SOAP Header
soap init(&soap);
soap.send timeout = 1; // 1s timeout
soap.connect flags = SO BROADCAST; // required for broadcast
soap default SOAP ENV Header(&soap, &header); // init SOAP Header
header.wsa MessageID = ”message ID”;
header.wsa To = ”server URL”;
header.wsa Action = ”server action”;
soap.header = &header; // bind the SOAP Header for transport
// Send the message over UDP to a broadcast address:
if (soap send ns echoString(&soap, ”soap.udp://...”, NULL, ”hello world!”))
soap print fault(&soap, stderr); // report error
soap destroy(&soap); // cleanup
soap end(&soap); // cleanup
soap done(&soap); // close connection (should not use soap struct after this)
17.4
Client-side Request-Response Unicast
This example assumes that the gSOAP header file includes the SOAP Header with WS-Addressing
elements and the ns echoString function discussed in Section 17.1
struct soap soap;
struct SOAP ENV Header header; // the SOAP Header
181
struct wsa EndpointReferenceType replyTo; // (anonymous) reply address
char *res; // server response
soap init(&soap);
soap.send timeout = 1; // 1s timeout
soap.recv timeout = 1; // 1s timeout
soap default SOAP ENV Header(&soap, &header); // init SOAP Header
soap default wsa EndpointReferenceType(&soap, &replyTo); // init reply address
replyTo.Address = ”http://schemas.xmlsoap.org/ws/2004/08/addressing/role/anonymous”;
header.wsa MessageID = ”message ID”;
header.wsa To = ”server URL”;
header.wsa Action = ”server action”;
header.wsa ReplyTo = &replyTo;
soap.header = &header; // bind the SOAP Header for transport
// Send and receive messages over UDP:
if (soap call ns echoString(&soap, ”soap.udp://...”, NULL, ”hello world!”, &res))
{
if (soap.error == SOAP EOF && soap.errnum == 0)
// Timeout: no response from server (message already delivered?)
else
soap print fault(&soap, stderr);
}
else
// UDP server response is stored in ’res’
// check SOAP header received, if applicable
check header(&soap.header);
soap destroy(&soap); // cleanup
soap end(&soap); // cleanup
soap done(&soap); // close connection (should not use soap struct after this)
17.5
Client-side Request-Response Multicast
This example is similar to the request-response unicast example discussed above, but uses a broadcast address and the SO BROADCAST socket option. Because we expect to receive multiple responses, we also need to use separate request-response messages to send one request and consume
multiple responses. In this example we defined a bcastString request and a bcastStringResponse response message, which are essentially declared as one-way messages in the header file:
//gsoap ns service method-header-part:
//gsoap ns service method-header-part:
//gsoap ns service method-header-part:
//gsoap ns service method-header-part:
int ns bcastString(char *str, void);
//gsoap ns service method-header-part:
//gsoap ns service method-header-part:
//gsoap ns service method-header-part:
//gsoap ns service method-header-part:
int ns bcastStringResponse(char *res,
bcastString
bcastString
bcastString
bcastString
wsa
wsa
wsa
wsa
MessageID
To
Action
ReplyTo
bcastStringResponse
bcastStringResponse
bcastStringResponse
bcastStringResponse
void);
wsa
wsa
wsa
wsa
MessageID
To
Action
RelatesTo
The cliend code includes a loop to receive response messages until a timeout occurs:
182
struct soap soap;
struct SOAP ENV Header header;
struct wsa EndpointReferenceType replyTo;
char *res;
soap init(&soap);
soap.connect flags = SO BROADCAST;
soap.send timeout = 1; // 1s timeout
soap.recv timeout = 1; // 1s timeout
soap default SOAP ENV Header(&soap, &header);
soap.header = &header;
soap default wsa EndpointReferenceType(&soap, &replyTo);
replyTo.Address = ”http://schemas.xmlsoap.org/ws/2004/08/addressing/role/anonymous”;
header.wsa MessageID = ”message ID”;
header.wsa To = ”server URL”;
header.wsa Action = ”server action”;
header.wsa ReplyTo = &replyTo;
if (soap send ns bcastString(&soap, ”soap.udp://...”, NULL, ”hello world!”))
soap print fault(&soap, stderr);
else
{
for (;;)
{
if (soap recv ns bcastStringResponse(&soap, &res))
break;
// Got response ’res’ from a server
}
if (soap.error == SOAP EOF && soap.errnum == 0)
// Timeout: no more messages received
else
soap print fault(&soap, stderr);
}
soap destroy(&soap); // cleanup
soap end(&soap); // cleanup
soap done(&soap); // close connection (should not use soap struct after this)
Note that a server for the bcastString does not need to use two-one way messages. Thus, multicast
request-response message pattern can be declared and implemented as request-response operations
at the server side.
17.6
SOAP-over-UDP Server
The following example code illustrates a SOAP-over-UDP server for one-way sendString and requestresponse echoString messages. This example assumes that the gSOAP header file includes the SOAP
Header with WS-Addressing elements and the ns echoString function discussed in Section 17.1.
main()
{
struct soap soap;
soap init1(&soap, SOAP IO UDP); // must set UDP flag
// bind to host (NULL=current host) and port:
183
if (!soap valid socket(soap bind(&soap, host, port, 100)))
{
soap print fault(&soap, stderr);
exit(1);
}
for (;;)
{
if (soap serve(&soap))
soap print fault(&soap, stderr); // report the problem
soap destroy(&soap);
soap end(&soap);
}
soap done(&soap); // close connection
}
int ns echoString(struct soap *soap, char *str, char **res)
{
if (!soap->header)
return soap sender fault(soap, ”No SOAP header”, NULL);
if (!soap->header->wsa MessageID)
return soap sender fault(soap, ”No WS-Addressing MessageID”, NULL);
soap->header->wsa RelatesTo = (struct wsa Relationship*)soap malloc(soap, sizeof(struct
wsa Relationship));
soap default wsa Relationship(soap, soap->header->wsa RelatesTo);
soap->header->wsa RelatesTo-> item = soap->header->wsa MessageID;
// must check for duplicate messages
if (check received(soap->header->wsa MessageID))
{
// Request message already received
return SOAP STOP; // don’t return response
}
if (!soap->header->wsa ReplyTo || !soap->header->wsa ReplyTo->Address)
return soap sender fault(soap, ”No WS-Addressing ReplyTo address”, NULL);
soap->header->wsa To = soap->header->wsa ReplyTo->Address;
soap->header->wsa MessageID = soap strdup(soap, soap int2s(soap, id count++)) ;
soap->header->wsa Action = ”http://genivia.com/udp/echoStringResponse”;
*res = str;
return SOAP OK;
}
int ns sendString(struct soap *soap, char *str)
{
if (!soap->header)
return SOAP STOP;
if (!soap->header->wsa MessageID)
return SOAP STOP;
// must check for duplicate messages
if (check received(soap->header->wsa MessageID))
return SOAP STOP;
return SOAP OK;
}
int ns sendStringResponse(struct soap *soap, char *res)
{ return SOAP NO METHOD; } // we don’t expect to serve this message
184
The server binds to a host and port and accepts messages in a tight sequential loop. Because
UDP does not have the equivalent of an accept the messages cannot be dispatched to threads, the
soap serve waits for a message and immediately accepts it. You can use a receive timeout to make
soap serve non-blocking.
18
18.1
Advanced Features
Internationalization
gSOAP uses regular strings by default. Regular strings cannot be used to hold UCS characters
outside of the character range [1,255]. gSOAP can handle wide-character content in two ways.
First, applications can utilize wide-character strings (wchar t*) instead of regular strings to store
wide-character content. For example, the xsd:string string schema type can be declared as a
wide-character string and used subsequently:
typedef wchar t *xsd string;
...
int ns myMethod(xsd string input, xsd string *output);
Second, regular strings can be used to hold wide-character content in UTF-8 format. This is
accomplished with the SOAP C UTFSTRING flag (for both input/output mode), see Section 8.12.
With this flag set, gSOAP will deserialize XML into regular strings in UTF-8 format. An application
is responsible for filling regular strings with UTF-8 content to ensure that strings can be correctly
serialized XML. Third, the SOAP C MBSTRING flag (for both input/output mode) can be used to
activate multibyte character support. Multibyte support depends on the locale settings for dealing
with extended natural language encodings.
Both regular strings and wide-character strings can be used together within an application. For
example, the following header file declaration introduces two string schema types:
typedef wchar t *xsd string;
typedef char *xsd string ; // trailing ’ ’ avoids name clash
...
int ns myMethod(xsd string input, xsd string *output);
The input string parameter is a wide-character string and the output string parameter is a regular
string. The regular string has UCS character content in the range [1,255] unless the SOAP C UTFSTRING
flag is set. With this flag, the string has UTF-8 encoded content.
Please consult the UTF-8 specification for details on the UTF-8 format. Note that the ASCII
character set [1-127] is a subset of UTF-8. Therefore, with the SOAP C UTFSTRING flag set, strings
may hold ASCII character data and UTF-8 extensions.
18.2
Customizing the WSDL and Namespace Mapping Table File Contents
With gSOAP Directives
A header file can be augmented with directives for the gSOAP Stub and Skeleton compiler to
automatically generate customized WSDL and namespace mapping tables contents. The WSDL
185
and namespace mapping table files do not need to be modified by hand (Sections 7.2.8 and 9.4).
In addition, the sample SOAP/XML request and response files generated by the compiler are valid
provided that XML Schema namespace information is added to the header file with directives so
that the gSOAP compiler can produce example SOAP/XML messages that are correctly namespace
qualified. These compiler directive are specified as //-comments. (Note: blanks can be used
anywhere in the directive, except between // and gsoap.)
Three directives are currently supported that can be used to specify details associated with namespace prefixes used by the remote method names in the header file. To specify the name of a Web
Service in the header file, use:
//gsoap namespace-prefix service name: service-name
where namespace-prefix is a namespace prefix used by identifiers in the header file and service-name
is the name of a Web Service (only required to create new Web Services). The name may be followed
by text up to the end of the line which is incorporated into the WSDL service documentation.
Alternatively, the service documentation can be provided with the directive below.
To specify the name of the WSDL definitions in the header file, use:
//gsoap namespace-prefix service definitions: definitions-name
where namespace-prefix is a namespace prefix used by identifiers in the header file and definitionsname is the name of the WSDL definitions. By default, the WSDL definitions name is the same as
the service name.
To specify the documentation of a Web Service in the header file, use:
//gsoap namespace-prefix service documentation: text
where namespace-prefix is a namespace prefix used by identifiers in the header file and text is the
documentation text up to the end of the line. The text is incorporated into the WSDL service
documentation.
To specify the portType of a Web Service in the header file, use:
//gsoap namespace-prefix service portType: portType-name
or just
//gsoap namespace-prefix service type: portType-name
where namespace-prefix is a namespace prefix used by identifiers in the header file and portTypename is the portType name of the WSDL service portType.
To specify the port name of a Web Service in the header file, use:
//gsoap namespace-prefix service portName: port-name
186
where namespace-prefix is a namespace prefix used by identifiers in the header file and port-name is
the name of the WSDL service port element. By default, the port name is the same as the service
name.
To specify the binding name of a Web Service in the header file, use:
//gsoap namespace-prefix service binding: binding-name
where namespace-prefix is a namespace prefix used by identifiers in the header file and binding-name
is the binding name of the WSDL service binding element. By default, the binding name is the
same as the service name.
To specify the binding’s transport protocol of a Web Service in the header file, use:
//gsoap namespace-prefix service transport: transport-URL
where namespace-prefix is a namespace prefix used by identifiers in the header file and transportURL is the URL of the transport protocol such as http://schemas.xmlsoap.org/soap/http for HTTP.
HTTP transport is assumed by default.
To specify the location (or port endpoint) of a Web Service in the header file, use:
//gsoap namespace-prefix service location: URL
or alternatively
//gsoap namespace-prefix service port: URL
where URL is the location of the Web Service (only required to create new Web Services). The
URL specifies the path to the service executable (so URL/service-executable is the actual location of
the executable when declared).
To specify the name of the executable of a Web Service in the header file, use:
//gsoap namespace-prefix service executable: executable-name
where executable-name is the name of the executable of the Web Service.
When doc/literal encoding is required for the entire service, the service encoding can be specified
in the header file as follows:
//gsoap namespace-prefix service encoding: literal
or when the SOAP-ENV:encodingStyle attribute is different from the SOAP 1.1/1.2 encoding style:
//gsoap namespace-prefix service encoding: encoding-style
To specify the namespace URI of a Web Service in the header file, use:
//gsoap namespace-prefix service namespace: namespace-URI
187
where namespace-URI is the URI associated with the namespace prefix.
In addition, the schema namespace URI can be specified in the header file:
//gsoap namespace-prefix schema namespace: namespace-URI
where namespace-URI is the schema URI associated with the namespace prefix. If present, it
defines the schema-part of the generated WSDL file and the URI in the namespace mapping table.
This declaration is useful when the service declares its own data types that need to be associated
with a namespace. Furthermore, the header file for client applications do not need the full service
details and the specification of the schema namespaces for namespace prefixes suffices.
The directive above specifies a new schema and the gSOAP compiler generates a schema files (.xsd)
file for the schema. An existing schema namespace URI can be imported with:
//gsoap namespace-prefix schema import: namespace-URI
where namespace-URI is the schema URI associated with the namespace prefix. gSOAP does
not produce XML Schema files for imported schemas and imports the schema namespaces in the
generated WSDL file.
A schema namespace URI can be imported from a location with:
//gsoap namespace-prefix schema namespace: namespace-URI
//gsoap namespace-prefix schema import: schema-location
The elementFormDefault and attributeFormDefault qualification of a schema can be defined with:
//gsoap namespace-prefix schema elementForm: qualified
//gsoap namespace-prefix schema attributeForm: qualified
or:
//gsoap namespace-prefix schema elementForm: unqualified
//gsoap namespace-prefix schema attributeForm: unqualified
A shortcut to define the default qualification of elements and attributes of a schema:
//gsoap namespace-prefix schema form: qualified
or:
//gsoap namespace-prefix schema form: unqualified
To document a data type, use:
//gsoap namespace-prefix schema type-documentation: type-name //text
where type-name is the unqualified name of the data type and text is a line of text terminated by a
newline. Do not use any XML reserved characters in text such as < and >. Use well-formed XML
and XHTML markup instead. For example:
188
//gsoap ns schema type-documentation: tdata stores <a href="transaction.html">transaction</a>
data
class ns tdata
{ ... }
To document a method, use:
//gsoap namespace-prefix service method-documentation: method-name //text
where method-name is the unqualified name of the method and text is a line of text terminated
by a newline. Do not use any XML reserved characters in text such as < and >. Use well-formed
XML and XHTML markup instead. For example:
//gsoap ns service method-documentation: getQuote returns a <i>stock quote</i>
int ns getQuote(char *symbol, float & result);
To specify the SOAPAction for a method, use:
//gsoap namespace-prefix service method-action: method-name action
where method-name is the unqualified name of the method and action is a quoted or non-quoted
string (spaces and blanks are not allowed). For example:
//gsoap ns service method-action: getQuote ””
int ns getQuote(char *symbol, float & result);
When document style is preferred for a particular service method, use:
//gsoap namespace-prefix service method-style: method-name document
When SOAP RPC encoding is required for a particular service method, use:
//gsoap namespace-prefix service method-encoding: method-name encoded
When literal encoding is required for a particular service method, use:
//gsoap namespace-prefix service method-encoding: method-name literal
or when the SOAP-ENV:encodingStyle attribute is different from the SOAP 1.1/1.2 encoding style,
use:
//gsoap namespace-prefix service method-encoding: method-name encoding-style
When SOAP RPC encoding is required for a particular service method response when the request
message is literal, use:
//gsoap namespace-prefix service method-response-encoding: method-name encoded
189
When literal encoding is required for a particular service method response when the request message
is encoded, use:
//gsoap namespace-prefix service method-response-encoding: method-name literal
or when the SOAP-ENV:encodingStyle attribute is different from the SOAP 1.1/1.2 encoding style,
use:
//gsoap namespace-prefix service method-response-encoding: method-name encoding-style
Note that the method-response-encoding is set to the value of method-encoding by default.
When header processing is required, each method declared in the WSDL should provide a binding
to the parts of the header that may appear as part of a method request message. Such a binding
is given by:
//gsoap namespace-prefix service method-header-part: method-name header-part
For example:
struct SOAP ENV Header
{
char *h transaction;
struct UserAuth *h authentication;
};
Suppose method ns login uses both header parts (at most), then this is declared as:
//gsoap ns service method-header-part: login h transaction
//gsoap ns service method-header-part: login h authentication
int ns login(...);
Suppose method ns search uses only the first header part, then this is declared as:
//gsoap ns service method-header-part: search h transaction
int ns search(...);
Note that the method name and header part names MUST be namespace qualified. The headers
MUST be present in all operations that declared the header parts.
To specify the header parts for the method input (method request message), use:
//gsoap namespace-prefix service method-input-header-part: method-name header-part
Similarly, to specify the header parts for the method output (method response message), use:
//gsoap namespace-prefix service method-output-header-part: method-name header-part
The declarations above only affect the WSDL. For example:
190
struct SOAP ENV Header
{
char *h transaction;
struct UserAuth *h authentication;
};
//gsoap ns service method-input-header-part: login h authentication
//gsoap ns service method-input-header-part: login h transaction
//gsoap ns service method-output-header-part: login h transaction
int ns login(...);
The headers MUST be present in all operations that declared the header parts.
To specify MIME attachments for the method input and output (method request and response
messages), use:
//gsoap namespace-prefix service method-mime-type: method-name mime-type
You can repeat this directive for all multipartRelated MIME attachments you want to associate
with the method.
To specify MIME attachments for the method input (method request message), use:
//gsoap namespace-prefix service method-input-mime-type: method-name mime-type
Similarly, to specify MIME attachments for the method output (method response message), use:
//gsoap namespace-prefix service method-output-mime-type: method-name mime-type
You can repeat these directives for all multipartRelated MIME attachments you want to associate
with the method.
18.2.1
Example
The use of directives is best illustrated with an example. The quotex.h header file of the quotex
example in the gSOAP distribution for Unix/Linux is:
//gsoap ns1 service namespace: urn:xmethods-delayed-quotes
int ns1 getQuote(char *symbol, float &result);
//gsoap ns2 service namespace: urn:xmethods-CurrencyExchange
int ns2 getRate(char *country1, char *country2, float &result);
//gsoap
//gsoap
//gsoap
//gsoap
//gsoap
int ns3
ns3 service name: quotex
ns3 service style: rpc
ns3 service encoding: encoded
ns3 service location: http://www.cs.fsu.edu/˜engelen
ns3 service namespace: urn:quotex
getQuote(char *symbol, char *country, float &result);
191
The quotex example is a new Web Service created by combining two existing Web Services: the
XMethods Delayed Stock Quote service and XMethods Currency Exchange service.
Namespace prefix ns3 is used for the new quotex Web Service with namespace URI urn:quotex, service
name quotex, and location http://www.cs.fsu.edu/˜engelen. Since the new Web Service invokes the
ns1 getQuote and ns2 getRate remote methods, the service namespaces of these methods are given.
The service names and locations of these methods are not given because they are only required for
setting up a new Web Service for these methods (but may also be provided in the header file for
documentation purposes). After invoking the gSOAP Stub and Skeleton compiler on the quotex.h
header file:
soapcpp2 quotex.h
the WSDL of the new quotex Web Service is saved as quotex.wsdl. Since the service name (quotex), location (http://www.cs.fsu.edu/˜engelen), and namespace URI (urn:quotex) were provided in the header
file, the generated WSDL file does not need to be changed by hand and can be published immediately together with the compiled Web Service installed as a CGI application at the designated
URL (http://www.cs.fsu.edu/˜engelen/quotex.cgi and http://www.cs.fsu.edu/˜engelen/quotex.wsdl).
The namespace mapping table for the quotex.cpp Web Service implementation is saved as quotex.nsmap. This file can be directly included in quotex.cpp instead of specified by hand in the source
of quotex.cpp:
#include ”quotex.nsmap”
The automatic generation and inclusion of the namespace mapping table requires compiler directives
for all namespace prefixes to associate each namespace prefix with a namespace URI. Otherwise,
namespace URIs have to be manually added to the table (they appear as http://tempuri.org).
18.3
Transient Data Types
There are situations when certain data types have to be ignored by gSOAP for the compilation
of (de)marshalling routines. For example, in certain cases only a few members of a class or struct
need not be (de)serialized, or the base class of a derived class should not be (de)serialized. Certain
built-in classes such as ostream cannot be (de)serialized. Data parts that should be kept invisible
to gSOAP are called “transient”. Transient data types and transient struct/class members are
declared with the extern keyword or are declared within [ and ] blocks in the header file. The extern
keyword has a special meaning to the gSOAP compiler and won’t affect the generated codes. The
special [ and ] block construct can be used with data type declarations and within struct and class
declarations. The use of extern or [ ] achieve the same effect, but [ ] may be more convenient to
encapsulate transient types in a larger part of the header file. The use of extern with typedef is
reserved for the declaration of user-defined external (de)serializers for data types, see Section 18.5.
First example:
extern class ostream; // ostream can’t be (de)serialized, but need to be declared to make it visible
to gSOAP
class ns myClass
192
{ ...
virtual void print(ostream &s) const; // need ostream here
...
};
Second example:
[
class myBase // base class need not be (de)serialized
{ ... };
]
class ns myDerived : myBase
{ ... };
Third example:
[ typedef int transientInt; ]
class ns myClass
{
int a; // will be (de)serialized
[
int b; // transient field
char s[256]; // transient field
]
extern float d; // transient field
char *t; // will be (de)serialized
transientInt *n; // transient field
[
virtual void method(char buf[1024]); // does not create a char[1024] (de)serializer
]
};
In this example, class ns myClass has three transient fields: b, s, and n which will not be (de)serialized
in SOAP. Field n is transient because the type is declared within a transient block. Pointers,
references, and arrays of transient types are transient. The single class method is encapsulated
within [ and ] to prevent gSOAP from creating (de)serializers for the char[1024] type. gSOAP will
generate (de)serializers for all types that are not declared within a [ and ] transient block.
18.4
Volatile Data Types
While transient data types are supposed to be hidden from gSOAP, volatile data types are visible to
gSOAP but their declaration and implementation is assumed to be hidden. That is, volatile data
types are assumed to be part of an existing non-modifiable software package, such as a built-in
library. It would not make sense to redefine the data types in a gSOAP header file. When you need
to (de)serialize such data types, you must declare them in a gSOAP header file and use the volatile
qualifier.
Consider for example struct tm, declared in time.h. The structure may actually vary between platforms, but the tm structure includes at least the following fields:
193
volatile struct tm
{
int tm sec; /* seconds (0 - 60) */
int tm min; /* minutes (0 - 59) */
int tm hour; /* hours (0 - 23) */
int tm mday; /* day of month (1 - 31) */
int tm mon; /* month of year (0 - 11) */
int tm year; /* year - 1900 */
int tm wday; /* day of week (Sunday = 0) */
int tm yday; /* day of year (0 - 365) */
int tm isdst; /* is summer time in effect? */
char *tm zone; /* abbreviation of timezone name */
long tm gmtoff; /* offset from UTC in seconds */
};
Note that we qualified the structure volatile in the gSOAP header file to inform the gSOAP compiler
that it should not attempt to redeclare it. We can now readily serialize and deserialize the tm
structure. The following program fragment serializes the local time stored in a tm structure to
stdout:
struct soap *soap = soap new();
...
time t T = time(NULL);
struct tm *t = localtime(&T);
struct soap *soap = soap new();
soap set omode(soap, SOAP XML GRAPH); // good habit to use this
soap begin send(soap);
soap put tm(soap, t, ”myLocalTime”, NULL);
soap end send(soap);
soap destroy(soap);
soap end(soap);
soap free(soap); // detach and free context
It is also possible to serialize the tm fields as XML attributes using the @ qualifier, see Section 10.6.7.
If you must produce a schema file, say time.xsd, that defines an XML schema and namespace for
the tm struct, you can add a typedef declaration to the header file:
typedef struct tm time struct tm;
We used the typedef name time struct tm rather than time tm, because a schema name clash will
occur with the latter since taking off the time prefix will result in the same name being used.
Classes should be declared volatile to prevent modification of these classes by gSOAP. gSOAP adds
serialization methods to classes to support polymorphism. However, this is a problem when you
can’t modify class declarations because they are part of a non-modifiable software package. The
solution is to declare these classes volatile, similar to the tm structure example illustrated above.
You can also use a typedef to associate a schema with a class.
194
18.5
How to Declare User-Defined Serializers and Deserializers
Users can declare their own (de)serializers for specific data types instead of relying on the gSOAPgenerated (de)serializers. To declare a external (de)serializer, declare a type with extern typedef .
gSOAP will not generate the (de)serializers for the type name that is declared. For example:
extern typedef char *MyData;
struct Sample
{
MyData s; // use user-defined (de)serializer for this field
char *t; // use gSOAP (de)serializer for this field
};
The user is required to supply the following routines for each extern typedef ’ed name T:
void soap serialize T(struct soap *soap, const T *a)
void soap default T(struct soap *soap, T *a)
void soap out T(struct soap *soap, const char *tag, int id, const T *a, const char *type)
T *soap in T(struct soap *soap, const char *tag, T *a, const char *type)
The function prototypes can be found in soapH.h.
For example, the (de)serialization of MyData can be done with the following code:
void soap serialize MyData(struct soap *soap, MyData *const*a)
{ } // no need to mark this node (for multi-ref and cycle detection)
void soap default MyData(&soap, MyData **a)
{ *a = NULL }
void soap out MyData(struct soap *soap, const char *tag, int id, MyData *const*a, const char
*type)
{
soap element begin out(soap, tag, id, type); // print XML beginning tag
soap send(soap, *a); // just print the string (no XML conversion)
soap element end out(soap, tag); // print XML ending tag
}
MyData **soap in MyData(struct soap *soap, const char *tag, MyData **a, const char *type)
{
if (soap element begin in(soap, tag))
return NULL;
if (!a)
a = (MyData**)soap malloc(soap, sizeof(MyData*));
if (soap->null)
*a = NULL; // xsi:nil element
if (*soap->type && soap match tag(soap, soap->type, type))
{
soap->error = SOAP TYPE;
return NULL; // type mismatch
}
if (*soap->href)
a = (MyData**)soap id forward(soap, soap->href, a, SOAP MyData, sizeof(MyData*))
else if (soap->body)
195
{
char *s = soap value(soap); // fill buffer
*a = (char*)soap malloc(soap, strlen(s)+1);
strcpy(*a, s);
}
if (soap->body && soap element end in(soap, tag))
return NULL;
return a;
More information on custom (de)serialization will be provided in this document or in a separate
document in the future. The writing of the (de)serializer code requires the use of the low-level
gSOAP API.
18.6
How to Serialize Data Without Generating XSD Type Attributes
gSOAP serializes data in XML with xsi:type attributes when the types are declared with namespace
prefixes to indicate the schema type of the data contained in the elements. SOAP 1.1 and 1.2
requires xsi:type attributes in the presence of polymorphic data or when the type of the data
cannot be deduced from the SOAP payload. The namespace prefixes are associated with the type
names of typedef s (Section 10.3) for primitive data types, struct/class names, and enum names.
To prevent the output of these xsi:type attributes in the XML serialization, you can simply use
type declarations that do not include these namespace prefixes. That is, don’t use the typedef s for
primitive types and use unqualified type names with structs, classes, and enums.
However, there are two issues. Firstly, if you want to use a primitive schema type that has no
C/C++ counterpart, you must declare it as a typedef name with a leading underscore, as in:
typedef char * xsd date;
This will produce the necessary xsd:date information in the WSDL output by the gSOAP compiler.
But the XML serialization of this type at run time won’t include the xsi:type attribute. Secondly,
to include the proper schema definitions in the WSDL produced by the gSOAP compiler, you
should use qualified struct, class, and enum names with a leading underscore, as in:
struct ns myStruct
{ ... };
This ensures that myStruct is associated with a schema, and therefore included in the appropriate
schema in the generated WSDL. The leading underscore prevents the XML serialization of xsi:type
attributes for this type in the SOAP/XML payload.
18.7
Function Callbacks for Customized I/O and HTTP Handling
gSOAP provides five callback functions for customized I/O and HTTP handling:
196
Callback (function pointer)
int (*soap.fopen)(struct soap *soap, const char *endpoint, const char *host, int port)
Called from a client proxy to open a connection to a Web Service located at endpoint. Input
parameters host and port are micro-parsed from endpoint. Should return a valid file descriptor, or
SOAP INVALID SOCKET and soap->error set to an error code. Built-in gSOAP function: tcp connect
int (*soap.fget)(struct soap *soap)
Called by the main server loop upon an HTTP GET request. The SOAP GET METHOD error is
returned by the default fget routine. This callback can be used to respond to HTTP GET methods
with content, see Section 18.10. Should return SOAP OK, or a gSOAP error code. Built-in gSOAP
function: http get
int (*soap.fform)(struct soap *soap)
Called by the main server loop when a user-defined fparsehdr callback returned SOAP FORM to
signal that the HTTP body must be processed by this form handler callback. The HTTP POST
form data MUST be read, otherwise keep-alive messages will end up out of sync. Should return
SOAP OK or a gSOAP error code. Built-in gSOAP function: none.
int (*soap.fpost)(struct soap *soap, const char *endpoint, const char *host, int port, const char
*path, const char *action, size t count)
Called from a client proxy to generate the HTTP header to connect to endpoint. Input parameters
host, port, and path are micro-parsed from endpoint, action is the SOAP action, and count is the
length of the SOAP message or 0 when SOAP ENC XML is set or when SOAP IO LENGTH is reset.
Use function soap send(struct soap *soap, char *s) to write the header contents. Should return
SOAP OK, or a gSOAP error code. Built-in gSOAP function: http post.
int (*soap.fposthdr)(struct soap *soap, const char *key, const char *val)
Called by http post and http response (through the callbacks). Emits HTTP key: val header entries.
Should return SOAP OK, or a gSOAP error code. Built-in gSOAP function: http post header.
int (*soap.fresponse)(struct soap *soap, int soap error code, size t count)
Called from a service to generate the response HTTP header. Input parameter soap error code
is a gSOAP error code (see Section 9.2 and count is the length of the SOAP message or 0 when
SOAP ENC XML is set or when SOAP IO LENGTH is reset. Use function soap send(struct soap
*soap, char *s) to write the header contents. Should return SOAP OK, or a gSOAP error code
Built-in gSOAP function: http response
int (*soap.fparse)(struct soap *soap)
Called by client proxy and service to parse an HTTP header (if present). When user-defined, this
routine must at least skip the header. Use function int soap getline(struct soap *soap, char *buf, int
len) to read HTTP header lines into a buffer buf of length len (returns empty line at end of HTTP
header). Should return SOAP OK, or a gSOAP error code. Built-in gSOAP function: http parse
int (*soap.fparsehdr)(struct soap *soap, const char *key, const char *val)
Called by http parse (through the fparse callback). Handles HTTP key: val header entries to set
gSOAP’s internals. Should return SOAP OK, SOAP STOP (see fstop) or a gSOAP error code.
Built-in gSOAP function: http parse header
int (*soap.fclose)(struct soap *soap)
Called by client proxy multiple times, to close a socket connection before a new socket connection
is established and at the end of communications when the SOAP IO KEEPALIVE flag is not set and
soap.keep alive6=0 (indicating that the other party supports keep alive). Should return SOAP OK,
or a gSOAP error code. Built-in gSOAP function: tcp disconnect
int (*soap.fsend)(struct soap *soap, const char *s, size t n)
Called for all send operations to emit contents of s of length n. Should return SOAP OK, or a
gSOAP error code. Built-in gSOAP function: fsend
size t (*soap.frecv)(struct soap *soap, char *s, size t n)
Called for all receive operations to fill buffer s of maximum length n. Should return the number of
bytes read or 0 in case of an error, e.g. EOF. Built-in gSOAP function: frecv
int (*soap.fignore)(struct soap *soap, const char *tag)
Called when an unknown XML element was encountered on the input. The tag parameter is
197
the offending XML element tag name. Should return SOAP OK, or a gSOAP error code such
as SOAP TAG MISMATCH to throw an exception. Built-in gSOAP function: none.
int (*soap.fconnect)(struct soap *soap, const char *endpoint, const char *host, int port)
When non-NULL, this callback is called for all client-to-server connect operations instead of the builtin socket connect code. Therefore, it can be used to override the built-in connection establishment.
Parameter endpoint contains the server endpoint URL, host the domain name or IP, and port the
port number. Should return SOAP OK, or a gSOAP error code. Built-in gSOAP function: none
int (*soap.faccept)(struct soap *soap, struct sockaddr *a, int *n)
Called by soap accept. This is a wrapper routine for accept. Should return a valid socket descriptor or
SOAP INVALID SOCKET and set soap->error to an error code. Built-in gSOAP function: tcp accept
In addition, a void*user field in the struct soap data structure is available to pass user-defined data
to the callbacks.
The following example uses I/O function callbacks for customized serialization of data into a buffer
and deserialization back into a datastructure:
char buf[10000]; // XML buffer
int len1 = 0; // #chars written
int len2 = 0; // #chars read
// mysend: put XML in buf[]
int mysend(struct soap *soap, const char *s, size t n)
{
if (len1 + n > sizeof(buf))
return SOAP EOF;
strcpy(buf + len1, s);
len1 += n;
return SOAP OK;
}
// myrecv: get XML from buf[]
size t myrecv(struct soap *soap, char *s, size t n)
{
if (len2 + n > len1)
n = len1 - len2;
strncpy(s, buf + len2, n);
len2 += n;
return n;
}
main()
{
struct soap soap;
struct ns person p;
soap init(&soap);
len1 = len2 = 0; // reset buffer pointers
p.name = ”John Doe”;
p.age = 25;
soap.fsend = mysend; // assign callback
soap.frecv = myrecv; // assign callback
soap begin(&soap);
soap set omode(&soap, SOAP XML GRAPH);
soap serialize ns person(&soap, &p);
soap put ns person(&soap, &p, ”ns:person”, NULL);
if (soap.error)
{
soap print fault(&soap, stdout);
exit(1);
}
soap end(&soap);
soap begin(&soap);
soap get ns person(&soap, &p, ”ns:person”, NULL);
if (soap.error)
{
soap print fault(&soap, stdout);
198
exit(1);
}
soap destroy(&soap);
soap end(&soap);
soap done(&soap); // disable callbacks
}
The soap done function can be called to reset the callback to the default internal gSOAP I/O and
HTTP handlers.
The following example illustrates customized I/O and (HTTP) header handling. The SOAP request
is saved to a file. The client proxy then reads the file contents as the service response. To perform
this trick, the service response has exactly the same structure as the request. This is declared by
the struct ns test output parameter part of the remote method declaration. This struct resembles
the service request (see the generated soapStub.h file created from the header file).
The header file is:
//gsoap ns service name: callback
//gsoap ns service namespace: urn:callback
struct ns person
{
char *name;
int age;
};
int ns test(struct ns person in, struct ns test &out);
The client program is:
#include ”soapH.h”
...
int myopen(struct soap *soap, const char *endpoint, const char *host, int port)
{
if (strncmp(endpoint, ”file:”, 5))
{
printf(”File name expected\n”);
return SOAP EOF;
}
if ((soap->sendfd = soap->recvfd = open(host, O RDWR|O CREAT, S IWUSR|S IRUSR)) < 0)
return SOAP EOF;
return SOAP OK;
}
void myclose(struct soap *soap)
{
if (soap->sendfd > 2) // still open?
close(soap->sendfd); // then close it
soap->recvfd = 0; // set back to stdin
soap->sendfd = 1; // set back to stdout
}
int mypost(struct soap *soap, const char *endpoint, const char *host, const char *path, const
char *action, size t count)
199
{
return soap send(soap, ”Custom-generated file\n”); // writes to soap->sendfd
}
int myparse(struct soap *soap)
{
char buf[256];
if (lseek(soap->recvfd, 0, SEEK SET) < 0 || soap getline(soap, buf, 256)) // go to begin and
skip custom header
return SOAP EOF;
return SOAP OK;
}
main()
{
struct soap soap;
struct ns test r;
struct ns person p;
soap init(&soap); // reset
p.name = ”John Doe”;
p.age = 99;
soap.fopen = myopen; // use custom open
soap.fpost = mypost; // use custom post
soap.fparse = myparse; // use custom response parser
soap.fclose = myclose; // use custom close
soap call ns test(&soap, ”file://test.xml”, ””, p, r);
if (soap.error)
{
soap print fault(&soap, stdout);
exit(1);
}
soap end(&soap);
soap init(&soap); // reset to default callbacks
}
SOAP 1.1 and 1.2 specify that XML elements may be ignored when present in a SOAP payload
on the receiving side. gSOAP ignores XML elements that are unknown, unless the XML attribute
mustUnderstand="true" is present in the XML element. It may be undesirable for elements to be
ignored when the outcome of the omission is uncertain. The soap.fignore callback can be set to a function that returns SOAP OK in case the element can be safely ignored, or SOAP MUSTUNDERSTAND
to throw an exception, or to perform some application-specific action. For example, to throw an
exception as soon as an unknown element is encountered on the input, use:
int myignore(struct soap *soap, const char *tag)
{
return SOAP MUSTUNDERSTAND; // never skip elements (secure)
}
...
soap.fignore = myignore;
soap call ns method(&soap, ...); // or soap serve(&soap);
To selectively throw an exception as soon as an unknown element is encountered but element ns:xyz
can be safely ignored, use:
200
int myignore(struct soap *soap, const char *tag)
{
if (soap match tag(soap, tag, ”ns:xyz”) != SOAP OK)
return SOAP MUSTUNDERSTAND;
return SOAP OK;
}
...
soap.fignore = myignore;
soap call ns method(&soap, ...); // or soap serve(&soap)
...
struct Namespace namespaces[] =
{
{”SOAP-ENV”, ”http://schemas.xmlsoap.org/soap/envelope/”},
{”SOAP-ENC”,”http://schemas.xmlsoap.org/soap/encoding/”},
{”xsi”, ”http://www.w3.org/2001/XMLSchema-instance”},
{”xsd”, ”http://www.w3.org/2001/XMLSchema”},
{”ns”, ”some-URI”}, // the namespace of element ns:xyz
{NULL, NULL}
Function soap match tag compares two tags. The third parameter may be a pattern where * is a
wildcard and - is a single character wildcard. So for example soap match tag(tag, ”ns:*”) will match
any element in namespace ns or when no namespace prefix is present in the XML message.
The callback can also be used to keep track of unknown elements in an internal data structure such
as a list:
struct Unknown
{
char *tag;
struct Unknown *next;
};
int myignore(struct soap *soap, const char *tag)
{
char *s = (char*)soap malloc(soap, strlen(tag)+1);
struct Unknown *u = (struct Unknown*)soap malloc(soap, sizeof(struct Unknown));
if (s && u)
{
strcpy(s, tag);
u->tag = s;
u->next = ulist;
ulist = u;
}
}
...
struct soap *soap;
struct Unknown *ulist = NULL;
soap init(&soap);
soap.fignore = myignore;
soap call ns method(&soap, ...); // or soap serve(&soap)
// print the list of unknown elements
soap end(&soap); // clean up
201
18.8
HTTP 1.0 and 1.1
gSOAP uses HTTP 1.1 by default. You can revert to HTTP 1.0 as follows:
struct soap soap;
soap init(&soap);
...
soap.http version = ”1.0”;
This sets the HTTP version and reconfigures the engine to revert to HTTP 1.0. Note that you
cannot use HTTP chunking with HTTP 1.0.
18.9
HTTP 307 Temporary Redirect Support
The client-side handling of HTTP 307 code ”Temporary Redirect” and any of the redirect codes
301, 302, and 303 are not automated in gSOAP. Client application developers may want to consider
adding a few lines of code to support redirects. It was decided not to automatically support redirects
for the following reasons:
• Redirecting a secure HTTPS address to a non-secure HTTP address via 307 creates a security
vulnerability.
• Cyclic redirects must be detected (e.g. allowing only a limited number of redirect levels).
• Redirecting HTTP POST will result in re-serialization and re-post of the entire SOAP request.
The SOAP request message must be re-posted in its entirity when re-issuing the SOAP
operation to a new address.
To implement client-side 307 redirect, add the following lines of code:
char *endpoint = NULL; // use default endpoint given in WSDL (or add another one here)
int n = 10; // max redirect count
while (n–)
{
if (soap call ns1 myMethod(soap, endpoint, ...))
{
if ((soap->error >= 301 && soap->error <= 303) || soap->error == 307)
endpoint = soap->endpoint; // endpoint from HTTP 301, 302, 303, 307 Location header
else
{ ... report and handle error
break;
}
}
else
break;
}
202
18.10
HTTP GET Support
A gSOAP server normally only grants HTTP (and HTTPS) POST requests. To support HTTP
(HTTPS) GET, you need to set the soap.fget callback. The callback is required to produce a
response to the request in textual form, such as a Web page or a SOAP/XML response.
The following example produces a Web page upon a HTTP GET request (e.g. from a browser):
struct soap *soap = soap new();
soap->fget = http get();
...
soap serve(soap);
...
int http get(struct soap *soap)
{
soap response(soap, SOAP HTML); // HTTP response header with text/html
soap send(soap, ”¡HTML¿My Web server is operational.¡/HTML¿”);
soap end send(soap);
return SOAP OK;
}
The example below produces a WSDL file upon a HTTP GET with path ?wsdl:
int http get(struct soap *soap)
{
char *s = strchr(soap->path, ’ ?’);
if (!s || strcmp(s, ”?wsdl”))
return SOAP GET METHOD;
fd = fopen(”myservice.wsdl”, ”rb”); // open WSDL file to copy
if (!fd)
return 404; // return HTTP not found error
soap->http content = ”text/xml”; // HTTP header with text/xml content
soap response(soap, SOAP FILE);
for (;;)
{
r = fread(soap->tmpbuf, 1, sizeof(soap->tmpbuf), fd);
if (!r)
break;
if (soap send raw(soap, soap->tmpbuf, r))
break; // can’t send, but little we can do about that
}
fclose(fd);
soap end send(soap);
return SOAP OK;
}
Using one-way SOAP/XML message, you can also return a SOAP/XML response:
int http get(struct soap *soap)
{
if ((soap->omode & SOAP IO) != SOAP IO CHUNK)
203
soap set omode(soap, SOAP IO STORE); // if not chunking we MUST buffer entire content
to determine content length
soap response(soap, SOAP OK);
return soap send ns1 mySendMethodResponse(soap, ””, NULL, ... params ...);
}
where ns1 mySendMethodResponse is a one-way message declared in a gSOAP header file as:
int ns1 mySendMethodResponse(... params ..., void);
The generated soapClient.cpp includes the sending-side stub function.
18.11
HTTP Keep-Alive
gSOAP supports keep-alive socket connections. To activate keep-alive support, set the SOAP IO KEEPALIVE
flag for both input and output modes, see Section 8.12. For example
struct soap soap;
soap init2(&soap, SOAP IO KEEPALIVE, SOAP IO KEEPALIVE);
When a client or a service communicates with another client or service that supports keep alive, the
attribute soap.keep alive will be set to 1, otherwise it is reset to 0 (indicating that the other party will
close the connection). The connection maybe terminated on either end before the communication
completed, for example when the server keep-alive connection has timed out. This generates a
”Broken Pipe” signal on Unix/Linux platforms. This signal can be caught with a signal handler:
signal(SIGPIPE, sigpipe handle);
where, for example:
void sigpipe handle(int x) { }
Alternatively, broken pipes can be kept silent by setting:
soap.socket flags = MSG NOSIGNAL;
This setting will not generate a sigpipe but read/write operations return SOAP EOF instead. Note
that Win32 systems do not support signals and lack the MSG NOSIGNAL flag. The sigpipe handling
and flags are not very portable.
A connection will be kept open only if the request contains an HTTP 1.0 header with ”Connection:
Keep-Alive” or an HTTP 1.1 header that does not contain ”Connection: close”. This means
that a gSOAP client method call should use ”http://” in the endpoint URL of the request to the
stand-alone service to ensure HTTP headers are used.
If the client does not close the connection, the server will wait forever when no recv timeout is
specified. In addition, other clients will be denied service as long as a client keeps the connection
to the server open. To prevent this from happening, the service should be multi-threaded such that
each thread handles the client connection:
204
int main(int argc, char **argv)
{
struct soap soap, *tsoap;
pthread t tid;
int m, s;
soap init2(&soap, SOAP IO KEEPALIVE, SOAP IO KEEPALIVE);
soap.max keep alive = 100; // at most 100 calls per keep-alive session
soap.accept timeout = 600; // optional: let server time out after ten minutes of inactivity
m = soap bind(&soap, NULL, 18000, BACKLOG); // use port 18000 on the current machine
if (m < 0)
{
soap print fault(&soap, stderr);
exit(1);
}
fprintf(stderr, "Socket connection successful %d\n", m);
for (count = 0; count >= 0; count++)
{
soap.socket flags = MSG NOSIGNAL; // use this
soap.accept flags = SO NOSIGPIPE; // or this to prevent sigpipe
s = soap accept(&soap);
if (s < 0)
{
if (soap.errnum)
soap print fault(&soap, stderr);
else
fprintf(stderr, "Server timed out\n"); // Assume timeout is long enough for threads to
complete serving requests
break;
}
fprintf(stderr, "Accepts socket %d connection from IP %d.%d.%d.%d\n", s, (int)(soap.ip>>24)&0xFF,
(int)(soap.ip>>16)&0xFF, (int)(soap.ip>>8)&0xFF, (int)soap.ip&0xFF);
tsoap = soap copy(&soap);
pthread create(&tid, NULL, (void*(*)(void*))process request, (void*)tsoap);
}
return 0;
}
void *process request(void *soap)
{
pthread detach(pthread self());
((struct soap*)soap)->recv timeout = 300; // Timeout after 5 minutes stall on recv
((struct soap*)soap)->send timeout = 60; // Timeout after 1 minute stall on send
soap serve((struct soap*)soap);
soap destroy((struct soap*)soap);
soap end((struct soap*)soap);
soap free((struct soap*)soap);
return NULL;
}
To prevent a malicious client from keeping a thread waiting forever by keeping the connection
open, timeouts are set in the process request routine. See Section 18.17 for more details on timeout
settings.
205
A gSOAP client call will automatically attempt to re-establish a connection to a server when the
server has terminated the connection for any reason. This way, a sequence of calls can be made
to the server while keeping the connection open. Client stubs will poll the server to check if
the connection is still open. When the connection was terminated by the server, the client will
automatically reconnect.
A client should reset SOAP IO KEEPALIVE just before the last call to a server to close the connection
after this last call. This will close the socket after the call and also informs the server to gracefully
close the connection.
18.12
HTTP Chunked Transfer Encoding
gSOAP supports HTTP chunked transfer encoding. Un-chunking of inbound messages takes place
automatically. Outbound messages are never chunked, except when the SOAP IO CHUNK flag is set
for the output mode. Most Web services, however, will not accept chunked inbound messages.
18.13
HTTP Buffered Sends
The entire outbound message can be stored to determine the HTTP content length rather than the
two-phase encoding used by gSOAP which requires a separate pass over the data to determine the
length of the outbound message. Setting the flag SOAP IO STORE for the output mode will buffer
the entire message. This can speed up the transmission of messages, depending on the content, but
may require significant storage space to hold the verbose XML message.
Zlib compressed transfers require buffering. The SOAP IO STORE flag is set when the SOAP ENC ZLIB
flag is set to send compressed messages. The use of chunking significantly reduces memory usage
and may speed up the transmission of compressed SOAP/XML messages. This is accomplished by
setting the SOAP IO CHUNK flag with SOAP ENC ZLIB for the output mode.
18.14
HTTP Authentication
HTTP authentication (basic) is enabled at the client-side by setting the soap.userid and soap.passwd
strings to a username and password, respectively. A server may request user authentication and
denies access (HTTP 401 error) when the client tries to connect without HTTP authentication (or
with the wrong authentication information).
Here is an example client code fragment to set the HTTP authentication username and password:
struct soap soap;
soap init(&soap);
soap.userid = ”guest”;
soap.passwd = ”visit”;
...
A client SOAP request will have the following HTTP header:
POST /XXX HTTP/1.0
Host: YYY
206
User-Agent: gSOAP/2.2
Content-Type: text/xml; charset=utf-8
Content-Length: nnn
Authorization: Basic Z3Vlc3Q6Z3Vlc3Q=
...
A client MUST set the soap.userid and soap.passwd strings for each call that requires client authentication. The strings are reset after each successful or unsuccessful call.
When present, the value of the WWW-Authenticate HTTP header with the authentication realm
can be obtained from the soap.authrealm string. This is useful for clients to respond intelligently to
authentication requests.
A stand-alone gSOAP Web Service can enforce HTTP authentication upon clients, by checking
the soap.userid and soap.passwd strings. These strings are set when a client request contains HTTP
authentication headers. The strings SHOULD be checked in each service method (that requires
authentication to execute).
Here is an example service method implementation that enforced client authentication:
int ns method(struct soap *soap, ...)
{
if (!soap->.userid || !soap->.passwd || strcmp(soap->.userid, ”guest”) || strcmp(soap->.passwd,
”visit”))
return 401; ...
}
When the authentication fails, the service response with a SOAP Fault message and a HTTP error
code ”401 Unauthorized”. The HTTP error codes are described in Section 9.2.
18.15
HTTP Proxy Authentication
HTTP proxy authentication (basic) is enabled at the client-side by setting the soap.proxy userid and
soap.proxy passwd strings to a username and password, respectively. For example, a proxy server may
request user authentication. Otherwise, access is denied by the proxy (HTTP 407 error). Example
client code fragment to set proxy server, username, and password:
struct soap soap;
soap init(&soap);
soap.proxy host = ”xx.xx.xx.xx”; // IP
soap.proxy port = 8080;
soap.proxy userid = ”guest”;
soap.proxy passwd = ”guest”;
...
A client SOAP request will have the following HTTP header:
POST /XXX HTTP/1.0
Host: YYY
User-Agent: gSOAP/2.2
Content-Type: text/xml; charset=utf-8
207
Content-Length: nnn
Proxy-Authorization:
...
18.16
Basic Z3Vlc3Q6Z3Vlc3Q=
Speed Improvement Tips
Here are some tips you can use to speed up gSOAP. gSOAP’s default settings are choosen to
maximize portability and compatibility. The settings can be tweaked to optimize the performance
as follows:
• Increase the buffer size SOAP BUFLEN by changing the SOAP BUFLEN macro in stdsoap2.h. Use
buffer size 21 8 = 262144 for example.
• Use HTTP keep-alive at the client-side, see 18.11, when the client needs to make a series
of calls to the same server. Server-side keep-alive support can greatly improve performance
of both client and server. But be aware that clients and services under Unix/Linux require
signal handlers to catch dropped connections.
• Use HTTP chunked transfers, see 18.12.
• Do NOT use gzip compression, even when transferring data over a modem connection.
Modems already compress data transfers.
18.17
Timeout Management for Non-Blocking Operations
Socket connect, accept, send, and receive timeout values can be set to manage socket communication timeouts. The soap.connect timeout, soap.accept timeout, soap.send timeout, and soap.recv timeout
attributes of the current gSOAP runtime environment soap can be set to the appropriate userdefined socket send, receive, and accept timeout values. A positive value measures the timeout in
seconds. A negative timeout value measures the timeout in microseconds (10−6 sec).
The soap.connect timeout specifies the timeout value for soap call ns method calls.
The soap.accept timeout specifies the timeout value for soap accept(&soap) calls.
The soap.send timeout and soap.recv timeout specify the timeout values for non-blocking socket I/O
operations.
Example:
struct soap soap;
soap init(&soap);
soap.send timeout = 10;
soap.recv timeout = 10;
This will result in a timeout if no data can be send in 10 seconds and no data is received within
10 seconds after initiating a send or receive operation over the socket. A value of zero disables
timeout, for example:
208
soap.send timeout = 0;
soap.recv timeout = 0;
When a timeout occurs in send/receive operations, a SOAP EOF exception will be raised (“end of
file or no input”). Negative timeout values measure timeouts in microseconds, for example:
#define uSec *-1
#define mSec *-1000
soap.accept timeout = 10 uSec;
soap.send timeout = 20 mSec;
soap.recv timeout = 20 mSec;
The macros improve readability.
Caution: Many Linux versions do not support non-blocking connect(). Therefore, setting soap.connect timeout
for non-blocking soap call ns method calls may not work under Linux.
18.18
Socket Options and Flags
gSOAP’s socket communications can be controlled with socket options and flags. The gSOAP
run-time environment struct soap flags are: int soap.socket flags to control socket send() and recv()
calls, int soap.connect flags to set client connection socket options, int soap.bind flags to set serverside port bind socket options, int soap.accept flags to set server-side request message accept socket
options. See the manual pages of send and recv for soap.socket flags values and see the manual pages
of setsockopt for soap.connect flags, soap.bind flags, and soap.accept flags (SOL SOCKET) values. These
SO socket option flags (see setsockopt manual pages) can be bit-wise or-ed to set multiple socket
options at once. The client-side flag soap.connect flags=SO LINGER is supported with values l onoff=1
and l linger=0.
For example, to disable sigpipe signals on Unix/Linux platforms use: soap.socket flags=MSG NOSIGNAL
and/or soap.connect flags=SO NOSIGPIPE (i.e. client-side connect) depending on your platform.
Use soap.bind flags=SO REUSEADDR to enable server-side port reuse and local port sharing (but be
aware of the possible security implications such as port hijacking).
Note that multiple socket options can be explicitly set with setsockopt as follows:
int sock = soap bind(soap, host, port, backlog);
if (soap valid socket(sock))
{
setsockopt(sock, ..., ..., ..., ...); setsockopt(sock, ..., ..., ..., ...);
18.19
Secure SOAP Web Services with HTTPS/SSL
When a Web Service is installed as CGI, it uses standard I/O that is encrypted/decrypted by the
Web server that runs the CGI application. Therefore, HTTPS/SSL support must be configured for
the Web server (not CGI-based Web Service application itself).
209
To enable OpenSSL, first install OpenSSL and use option -DWITH OPENSSL to compile the sources
with your C or C++ compiler, for example:
g++ -DWITH OPENSSL -o myprog myprog.cpp stdsoap2.cpp soapC.cpp soapServer.cpp -lssl lcrypto
SSL support for stand-alone gSOAP Web services is enabled by calling soap ssl accept after soap accept.
In addition, a key file, a CA file (or path to certificates), DH file (if RSA is not used), and password
need to be supplied. Instructions on how to do this can be found in the OpenSSL documentation
http://www.openssl.org. See also Section 18.22.
Let’s take a look at an example SSL secure multi-threaded stand-alone SOAP Web Service:
int main()
{
int m, s;
pthread t tid;
struct soap soap, *tsoap;
soap ssl init(); /* init OpenSSL (just once) */
if (CRYPTO thread setup())
{
fprintf(stderr, ”Cannot setup thread mutex\n”);
exit(1);
}
soap init(&soap);
if (soap ssl server context(&soap,
SOAP SSL DEFAULT,
”server.pem”, /* keyfile: required when server must authenticate to clients (see SSL docs on
how to obtain this file) */
”password”, /* password to read the key file */
”cacert.pem”, /* optional cacert file to store trusted certificates */
NULL, /* optional capath to directory with trusted certificates */
”dh512.pem”, /* DH file, if NULL use RSA */
NULL, /* if randfile!=NULL: use a file with random data to seed randomness */
NULL /* optional server identification to enable SSL session cache (must be a unique name)
*/ ))
{
soap print fault(&soap, stderr);
exit(1);
}
m = soap bind(&soap, NULL, 18000, 100); // use port 18000
if (m < 0)
{
soap print fault(&soap, stderr);
exit(1);
}
fprintf(stderr, ”Socket connection successful: master socket = %d\n”, m);
for (;;)
{
s = soap accept(&soap);
fprintf(stderr, ”Socket connection successful: slave socket = %d\n”, s);
210
if (s < 0)
{
soap print fault(&soap, stderr);
break;
}
tsoap = soap copy(&soap); /* should call soap ssl accept on a copy */
if (!tsoap)
break;
if (soap ssl accept(tsoap))
{
soap print fault(tsoap, stderr);
soap free(tsoap);
continue; /* when soap ssl accept fails, we should just go on */
}
pthread create(&tid, NULL, &process request, (void*)tsoap);
}
soap done(&soap); /* deallocates SSL context */
CRYPTO thread cleanup();
return 0;
}
void *process request(void *soap)
{
pthread detach(pthread self());
soap serve((struct soap*)soap);
soap destroy((struct soap*)soap);
soap end((struct soap*)soap);
soap done((struct soap*)soap);
free(soap);
return NULL;
}
The soap ssl server context function initializes the server-side SSL context. The server.pem key file
is the server’s private key. The cacert.pem is used to authenticate clients and contains the client
certificates. Alternatively a directory name can be specified. This directory is assumed to contain
the certificates. The dh512.pem file specifies that DH will be used for key agreement instead of RSA.
The randfile entry can be used to seed the PRNG. The last entry enable server-side session caching.
A unique server name is required.
The CRYPTO thread setup() and CRYPTO thread cleanup() routines can be found in openssl/crypto/threads/thlock.c. These routines are required to setup locks for multi-threaded applications that use SSL. We
give a Windows and POSIX threads implementation of these here:
#include <unistd.h> /* defines POSIX THREADS if pthreads are available */
#ifdef POSIX THREADS
# include <pthread.h>
#endif
#if defined(WIN32)
# define MUTEX TYPE HANDLE
# define MUTEX SETUP(x) (x) = CreateMutex(NULL, FALSE, NULL)
# define MUTEX CLEANUP(x) CloseHandle(x)
# define MUTEX LOCK(x) WaitForSingleObject((x), INFINITE)
211
# define MUTEX UNLOCK(x) ReleaseMutex(x)
# define THREAD ID GetCurrentThreadID()
#elif defined( POSIX THREADS)
# define MUTEX TYPE pthread mutex t
# define MUTEX SETUP(x) pthread mutex init(&(x), NULL)
# define MUTEX CLEANUP(x) pthread mutex destroy(&(x))
# define MUTEX LOCK(x) pthread mutex lock(&(x))
# define MUTEX UNLOCK(x) pthread mutex unlock(&(x))
# define THREAD ID pthread self()
#else
# error ”You must define mutex operations appropriate for your platform”
# error ”See OpenSSL /threads/th-lock.c on how to implement mutex on your platform”
#endif
struct CRYPTO dynlock value { MUTEX TYPE mutex; };
static MUTEX TYPE *mutex buf;
static struct CRYPTO dynlock value *dyn create function(const char *file, int line)
{
struct CRYPTO dynlock value *value;
value = (struct CRYPTO dynlock value*)malloc(sizeof(struct CRYPTO dynlock value));
if (value)
MUTEX SETUP(value->mutex);
return value;
}
static void dyn lock function(int mode, struct CRYPTO dynlock value *l, const char *file, int
line)
{
if (mode & CRYPTO LOCK)
MUTEX LOCK(l->mutex);
else
MUTEX UNLOCK(l->mutex);
}
static void dyn destroy function(struct CRYPTO dynlock value *l, const char *file, int line)
{
MUTEX CLEANUP(l-¿mutex);
free(l);
}
void locking function(int mode, int n, const char *file, int line)
{
if (mode & CRYPTO LOCK)
MUTEX LOCK(mutex buf[n]);
else
MUTEX UNLOCK(mutex buf[n]);
}
unsigned long id function()
{
return (unsigned long)THREAD ID;
}
int CRYPTO thread setup()
{
int i;
mutex buf = (MUTEX TYPE*)malloc(CRYPTO num locks() * sizeof(MUTEX TYPE));
212
if (!mutex buf)
return SOAP EOM;
for (i = 0; i < CRYPTO num locks(); i++)
MUTEX SETUP(mutex buf[i]);
CRYPTO set id callback(id function);
CRYPTO set locking callback(locking function);
CRYPTO set dynlock create callback(dyn create function);
CRYPTO set dynlock lock callback(dyn lock function);
CRYPTO set dynlock destroy callback(dyn destroy function);
return SOAP OK;
}
void CRYPTO thread cleanup()
{
int i;
if (!mutex buf)
return;
CRYPTO set id callback(NULL);
CRYPTO set locking callback(NULL);
CRYPTO set dynlock create callback(NULL);
CRYPTO set dynlock lock callback(NULL);
CRYPTO set dynlock destroy callback(NULL);
for (i = 0; i < CRYPTO num locks(); i++)
MUTEX CLEANUP(mutex buf[i]);
free(mutex buf);
mutex buf = NULL;
}
For Unix and Linux, make sure you have signal handlers set in your service and/or client applications
to catch broken connections (SIGPIPE):
signal(SIGPIPE, sigpipe handle);
where, for example:
void sigpipe handle(int x) { }
By default, clients are not required to authenticate. To support client authentication use the
following:
if (soap ssl server context(&soap,
SOAP SSL REQUIRE CLIENT AUTHENTICATION,
”server.pem”,
”password”,
”cacert.pem”,
NULL,
”dh512.pem”,
NULL,
NULL
))
{
213
soap print fault(&soap, stderr);
exit(1);
}
This requires each client to authenticate with its certificate.
The cacert file and capath are optional. Either one can be specified when clients must run on nontrusted systems. We want to avoid storing trusted certificates in the default location on the file
system when that is not secure. Therefore, a flat cacert.pem file or directory can be specified to
store trusted certificates.
The gSOAP distribution includes a cacerts.pem file with the certificates of all certificate authorities
such as Verisign. You can use this file to verify the authentication of servers that provide certificates
issued by these CAs.
The cacert.pem, client.pem, and server.pem files in the gSOAP distribution are examples of self-signed
certificates.
Caution: it is important that the WITH OPENSSL macro MUST be consistently defined to compile
the sources, such as stdsoap2.cpp, soapC.cpp, soapClient.cpp, soapServer.cpp, and all application sources
that include stdsoap2.h or soapH.h. If the macros are not consistently used, the application will crash
due to a mismatches in the declaration and access of the gSOAP environment.
18.20
Secure SOAP Clients with HTTPS/SSL
You need to install the OpenSSL library on your platform to enable secure SOAP clients to
utilize HTTPS/SSL. After installation, compile all the sources of your application with option
-DWITH OPENSSL. For example on Linux:
g++ -DWITH OPENSSL myclient.cpp stdsoap.cpp soapC.cpp soapClient.cpp -lssl -lcrypto
or Unix:
g++ -DWITH OPENSSL myclient.cpp stdsoap.cpp soapC.cpp soapClient.cpp -lxnet -lsocket -lnsl
-lssl -lcrypto
or you can add the following line to soapdefs.h:
#define WITH OPENSSL
and compile with option -DWITH SOAPDEFS H to include soapdefs.h in your project. A client program
simply uses the prefix https: instead of http: in the endpoint URL of a remote method call to a
Web Service to use encrypted transfers (if the service supports HTTPS). You need to specify the
client-side key file and password of the keyfile:
soap ssl init(); /* init OpenSSL (just once) */
if (soap ssl client context(&soap,
SOAP SSL DEFAULT,
”client.pem”, /* keyfile: required only when client must authenticate to server (see SSL docs on
214
how to obtain this file) */
”password”, /* password to read the key file */
”cacerts.pem”, /* cacert file to store trusted certificates (needed to verify server) */
capath to direcoty with trusted certificates */
NULL /* if randfile!=NULL: use a file with random data to seed randomness */
))
{
soap print fault(&soap, stderr);
exit(1);
}
soap call ns mymethod(&soap, ”https://domain/path/secure.cgi”, ””, ...);
NULL, /*
By default, server authentication is enabled and the cacerts.pem or capath must be set so that the CA
certificates of the server(s) are accessible at run time. The cacert.pem file included in the package
contains the certificates of common CAs. This file must be supplied with the client, if server
authentication is required. Althernatively, you can use the plugin/cacerts.h and plugin/cacerts.c code
to embed CA certificates in your client code.
For systems based on Microsoft windows, the WinInet module can be used instead, see mod gsoap/gsoap win/winine
To disable server authentication for testing purposes, use the following:
if (soap ssl client context(&soap,
SOAP SSL NO AUTHENTICATION,
NULL,
NULL,
NULL,
NULL,
NULL
))
{
soap print fault(&soap, stderr);
exit(1);
}
This also assumes that the server does not require clients to authenticate (the keyfile is absent).
Make sure you have signal handlers set in your application to catch broken connections (SIGPIPE):
signal(SIGPIPE, sigpipe handle);
where, for example:
void sigpipe handle(int x) { }
Caution: it is important that the WITH OPENSSL macro MUST be consistently defined to compile
the sources, such as stdsoap2.cpp, soapC.cpp, soapClient.cpp, soapServer.cpp, and all application sources
that include stdsoap2.h or soapH.h. If the macros are not consistently used, the application will
crash due to a mismatches in the declaration and access of the gSOAP environment. Caution:
concurrent client calls MUST be made using separate soap structs copied with soap copy from an
215
originating struct initialized with soap ssl client context. In addition, the thread initialization code
discussed in Section 18.19 MUST be used to properly setup OpenSSL in a multi-threaded client
application.
18.21
SSL Authentication Callback
gSOAP provides a callback function for authentication initialization:
Callback (function pointer)
int (*soap.fsslauth)(struct soap *soap)
Initialize the authentication information for clients and services, such as the certificate chain, password, read the key and/or DH file, generate an RSA key, and initialization of the RNG. Should
return a gSOAP error code or SOAP OK. Built-in gSOAP function: ssl auth init
18.22
SSL Certificates
The gSOAP distribution includes a cacerts.pem file with the certificates of all certificate authorities
(such as Verisign). You can use this file to verify the authentication of servers that provide certificates issued by these CAs. Just set the cafile parameter to the location of this file on your file
system. Therefore, when you obtain a certifice signed by a trusted CA such as Verisign, you can
simply use the cacerts.pem file to develop client applications that can verify the authenticity of your
server.
Althernatively, you can use the plugin/cacerts.h and plugin/cacerts.c code to embed CA certificates in
your client code.
For systems based on Microsoft windows, the WinInet module can be used instead, see mod gsoap/gsoap win/winine
The other .pem files in the gSOAP distribution are examples of self-signed certificates for testing
purposes (cacert.pem, client.pem, server.pem).
You can also create your own self-signed certificates. There is more than one way to generate the
necessary files for clients and servers. See http://www.openssl.org for information on OpenSSL and
http://sial.org/howto/openssl/ca/ on how to setup and manage a local CA and http://sial.org/howto/openssl/selfsigned/ on how to setup self-signed test certificates.
It is also possible to convert IIS-generated certificates to PEM format, see http://www.icewarp.com/Knowledgebase/6
for a discussion and examples.
Here is the simplest way to setup self-signed certificates. First you need to create a private Certificate Authority (CA). The CA is used in SSL to verify the authenticity of a given certificate.
The CA acts as a trusted third party who has authenticated the user of the signed certificate as
being who they say. The certificate is signed by the CA, and if the client trusts the CA, it will
trust your certificate. For use within your organization, a private CA will probably serve your
needs. However, if you intend use your certificates for a public service, you should probably obtain
a certificate from a known CA (e.g. VeriSign). In addition to identification, your certificate is also
used for encryption.
Creating certificates should be done through a CA to obtain signed certificates. But you can create
your own certificates for testing purposes as follows.
216
• Go to the OpenSSL bin directory (/usr/local/ssl by default and /System/Library/OpenSSL on
Mac OS X)
• There should be a file called openssl.cnf
• Create a new directory in your home account, e.g. $HOME/CA, and copy the openssl.cnf file
to this directory
• Modify openssl.cnf by changing the ’dir’ value to HOME/CA
• Copy the README.txt, root.sh, and cert.sh scripts from the gSOAP distribution package
located in the samples/ssl directory to HOME/CA
• Follow the README.txt instructions
You now have a self-signed CA root certificate cacert.pem and a server.pem (or client.pem) certificate in PEM format. The cacert.pem certificate is used in the cafile parameter of the soap ssl client context
(or soap ssl server context) at the client (or server) side to verify the authenticity of the peer. You can
also provide a capath parameter to these trusted certificates. The server.pem (or client.pem) must
be provided with the soap ssl server context at the server side (or soap ssl client context at the client
side) together with the password you entered when generating the certificate using cert.sh to access
the file. These certificates must be present to grant authentication requests by peers. In addition,
the server.pem (and client.pem) include the host name of the machine on which the application
runs (e.g. localhost), so you need to generate new certificates when migrating a server (or client).
Finally you need to generate Diffie-Helmann parameters for the server if you don’t want to use
RSA. Do the following at the prompt:
openssl dhparam -outform PEM -out dh.pem 512
File dh512.pem is the output file and 512 is the number of bits used.
18.23
SSL Hardware Acceleration
You can specify a hardware engine to enable hardware support for cryptographic acceleration. This
can be done once in a server or client with the following statements:
static const char *engine = ”cswift”; /* engine name */
int main()
{
...
ENGINE *e;
if (!(e = ENGINE by id(engine)))
fprintf(stderr, ”Error finding engine %s\n”, engine);
else if (!ENGINE set default(e, ENGINE METHOD ALL))
fprintf(stderr, ”Error using engine %s\n”, engine);
...
The following table lists the names of the hardware and software engines:
217
Name
openssl
openbsd dev crypto
cswift
chil
atalla
nuron
ubsec
aep
sureware
18.24
Description
The default software engine for cryptographic operations
OpenBSD supports kernel level cryptography
CryptoSwift acceleration hardware
nCipher CHIL acceleration hardware
Compaq Atalla acceleration hardware
Nuron acceleration hardware
Broadcom uBSec acceleration hardware
Aep acceleration hardware
SureWare acceleration hardware
SSL on Windows
Set the full path to libssl.lib and libcrypto.lib under the MSVC++ ”Projects” menu, then choose
”Link”: ”Object/Modules”. The path to libssl32.dll and libeay32.dll need to be specified in the
PATH environment variable when running gSOAP applications.
Alternatively, you can use the WinInet interface available in the mod gsoap directory of the gSOAP
package. API instructions are included in the source.
18.25
Zlib Compression
To enable deflate and gzip compression with Zlib, install Zlib from http://www.zlib.org if not
already installed on your system. Compile stdsoap2.cpp (or stdsoap2.c) and all your sources that
include stdsoap2.h or soapH.h with compiler option -DWITH GZIP and link your code with the Zlib
library, e.g. -lz on Unix/Linux platforms.
The gzip compression is orthogonal to all transport encodings such as HTTP, SSL, DIME, and can
be used with other transport layers. You can even save and load compressed XML data to/from
files.
gSOAP supports two compression formats: deflate and gzip. The gzip format is used by default.
The gzip format has several benefits over deflate. Firstly, gSOAP can automatically detect gzip
compressed inbound messages, even without HTTP headers, by checking for the presence of a gzip
header in the message content. Secondly, gzip includes a CRC32 checksum to ensure messages
have been correctly received. Thirdly, gzip compressed content can be decompressed with other
compression software, so you can decompress XML data saved by gSOAP in gzip format.
Gzip compression is enabled by compiling the sources with -DWITH GZIP. To transmit gzip compressed SOAP/XML data, set the output mode flags to SOAP ENC ZLIB. For example:
soap init(&soap);
...
soap set omode(&soap, SOAP ENC ZLIB); // enable Zlib’s gzip
if (soap call ns myMethod(&soap, . . . ))
...
soap clr omode(&soap, SOAP ENC ZLIB); // disable Zlib’s gzip
...
218
This will send a compressed SOAP/XML request to a service, provided that Zlib is installed
and linked with the application and the -DWITH GZIP option was used to compile the sources.
Receiving compressed SOAP/XML over HTTP either in gzip or deflate formats is automatic. The
SOAP ENC ZLIB flag does not have to be set at the server side to accept compressed messages.
Reading and receiving gzip compressed SOAP/XML without HTTP headers (e.g. with other
transport protocols) is also automatic.
To control the level of compression for outbound messages, you can set the soap.z level to a value
between 1 and 9, where 1 is the best speed and 9 is the best compression (default is 6). For example
soap init(&soap);
...
soap set omode(&soap, SOAP ENC ZLIB);
soap.z level = 9; // best compression
...
To verify and monitor compression rates, you can use the values soap.z ratio in and soap.z ratio out.
These two float values lie between 0.0 and 1.0 and express the ratio of the compressed message
length over uncompressed message length.
soap call ns myMethod(&soap, . . . );
...
printf(”Compression ratio: %f%% (in) %f%% (out)\n”, 100*soap.z ratio out, 100*soap.z ratio in);
...
Note: lower ratios mean higher compression rates.
Compressed transfers require buffering the entire output message to determine HTTP message
length. This means that the SOAP IO STORE flag is automatically set when the SOAP ENC ZLIB flag
is set to send compressed messages. The use of HTTP chunking significantly reduces memory usage
and may speed up the transmission of compressed SOAP/XML messages. This is accomplished by
setting the SOAP IO CHUNK flag with SOAP ENC ZLIB for the output mode. However, some Web
servers do not accept HTTP chunked request messages (even when they return HTTP chunked
messages!). Stand-alone gSOAP services always accept chunked request messages.
To restrict the compression to the deflate format only, compile the sources with -DWITH ZLIB. This
limits compression and decompression to the deflate format. Only plain and deflated messages
can be exchanged, gzip is not supported with this option. Receiving gzip compressed content is
automatic, even in the absence of HTTP headers. Receiving deflate compressed content is not
automatic in the absence of HTTP headers and requires the flag SOAP ENC ZLIB to be set for the
input mode to decompress deflated data.
Caution: it is important that the WITH GZIP and WITH ZLIB macros MUST be consistently defined to compile the sources, such as stdsoap2.cpp, soapC.cpp, soapClient.cpp, soapServer.cpp, and all
application sources that include stdsoap2.h or soapH.h. If the macros are not consistently used, the
application will crash due to a mismatches in the declaration and access of the gSOAP environment.
219
18.26
Client-Side Cookie Support
Client-side cookie support is optional. To enable cookie support, compile all sources with option
-DWITH COOKIES, for example:
g++ -DWITH COOKIES -o myclient stdsoap2.cpp soapC.cpp soapClient.cpp
or add the following line to stdsoap.h:
#define WITH COOKIES
Client-side cookie support is fully automatic. So just (re)compile stdsoap2.cpp with -DWITH COOKIES
to enable cookie-based session control in your client.
A database of cookies is kept and returned to the appropriate servers. Cookies are not automatically saved to a file by a client. An example cookie file manager is included as an extras in the
distribution. You should explicitly remove all cookies before terminating a gSOAP environment by
calling soap free cookies(soap) or by calling soap done(soap).
To avoid ”cookie storms” caused by malicious servers that return an unreasonable amount of
cookies, gSOAP clients/servers are restricted to a database size that the user can limit (32 cookies
by default), for example:
struct soap soap;
soap init(&soap);
soap.cookie max = 10;
The cookie database is a linked list pointed to by soap.cookies where each node is declared as:
struct soap cookie
{
char *name;
char *value;
char *domain;
char *path;
long expire; /* client-side: local time to expire; server-side: seconds to expire */
unsigned int version;
short secure;
short session; /* server-side */
short env; /* server-side: 1 = got cookie from client */
short modified; /* server-side: 1 = client cookie was modified */
struct soap cookie *next;
};
Since the cookie database is linked to a soap struct, each thread has a local cookie database in a
multi-threaded implementation.
18.27
Server-Side Cookie Support
Server-side cookie support is optional. To enable cookie support, compile all sources with option
-DWITH COOKIES, for example:
220
g++ -DWITH COOKIES -o myserver ...
gSOAP provides the following cookie API for server-side cookie session control:
Function
struct soap cookie *soap set cookie(struct soap *soap, const char *name, const char *value, const
char *domain, const char *path);
Add a cookie to the database with name name and value value. domain and path may be NULL to
use the current domain and path given by soap cookie domain and soap cookie path. If successful,
returns pointer to a cookie node in the linked list, or NULL otherwise.
struct soap cookie *soap cookie(struct soap *soap, const char *name, const char *domain, const
char *path);
Find a cookie in the database with name name and value value. domain and path may be NULL to
use the current domain and path given by soap cookie domain and soap cookie path. If successful,
returns pointer to a cookie node in the linked list, or NULL otherwise.
char *soap cookie value(struct soap *soap, const char *name, const char *domain, const char *path);
Get value of a cookie in the database with name name. domain and path may be NULL to use the
current domain and path given by soap cookie domain and soap cookie path. If successful, returns
the string pointer to the value, or NULL otherwise.
long soap cookie expire(struct soap *soap, const char *name, const char *domain, const char *path);
Get expiration value of the cookie in the database with name name (in seconds). domain and path
may be NULL to use the current domain and path given by soap cookie domain and soap cookie path.
Returns the expiration value, or -1 if cookie does not exist.
int soap set cookie expire(struct soap *soap, const char *name, long expire, const char *domain,
const char *path);
Set expiration value expire of the cookie in the database with name name (in seconds). domain
and path may be NULL to use the current domain and path given by soap cookie domain and
soap cookie path. If successful, returns SOAP OK, or SOAP EOF otherwise.
int soap set cookie session(struct soap *soap, const char *name, const char *domain, const char
*path);
Set cookie in the database with name name to be a session cookie. This means that the cookie will be
returned to the client. (Only cookies that are modified are returned to the client). domain and path
may be NULL to use the current domain and path given by soap cookie domain and soap cookie path.
If successful, returns SOAP OK, or SOAP EOF otherwise.
int soap clr cookie session(struct soap *soap, const char *name, const char *domain, const char
*path);
Clear cookie in the database with name name to be a session cookie. domain and path may be NULL
to use the current domain and path given by soap cookie domain and soap cookie path. If successful,
returns SOAP OK, or SOAP EOF otherwise.
void soap clr cookie(struct soap *soap, const char *name, const char *domain, const char *path);
Remove cookie from the database with name name. domain and path may be NULL to use the
current domain and path given by soap cookie domain and soap cookie path.
int soap getenv cookies(struct soap *soap);
Initializes cookie database by reading the ’HTTP COOKIE’ environment variable. This provides a
means for a CGI application to read cookies send by a client. If successful, returns SOAP OK, or
SOAP EOF otherwise.
void soap free cookies(struct soap *soap);
Release cookie database.
The following global variables are used to define the current domain and path:
221
Attribute
const char *cookie domain
const char *cookie path
int cookie max
value
MUST be set to the domain (host) of the service
MAY be set to the default path to the service
maximum cookie database size (default=32)
The cookie path value is used to filter cookies intended for this service according to the path prefix
rules outlined in RFC2109.
The following example server adopts cookies for session control:
int main()
{
struct soap soap;
int m, s;
soap init(&soap);
soap.cookie domain = ”...”;
soap.cookie path = ”/”; // the path which is used to filter/set cookies with this destination
if (argc < 2)
{
soap getenv cookies(&soap); // CGI app: grab cookies from ’HTTP COOKIE’ env var
soap serve(&soap);
}
else
{
m = soap bind(&soap, NULL, atoi(argv[1]), 100);
if (m < 0)
exit(1);
for (int i = 1; ; i++)
{
s = soap accept(&soap);
if (s < 0)
exit(1);
soap serve(&soap);
soap end(&soap); // clean up
soap free cookies(&soap); // remove all old cookies from database so no interference occurs
with the arrival of new cookies
}
}
return 0;
}
int ck demo(struct soap *soap, ...)
{
int n;
const char *s;
s = soap cookie value(soap, ”demo”, NULL, NULL); // cookie returned by client?
if (!s)
s = ”init-value”; // no: set initial cookie value
else
... // modify ’s’ to reflect session control
soap set cookie(soap, ”demo”, s, NULL, NULL);
soap set cookie expire(soap, ”demo”, 5, NULL, NULL); // cookie may expire at client-side in 5
seconds
222
return SOAP OK;
}
18.28
Connecting Clients Through Proxy Servers
When a client needs to connect to a Web Service through a proxy server, set the soap.proxy host
string and soap.proxy port integer attributes of the current soap runtime environment to the proxy’s
host name and port, respectively. For example:
struct soap soap;
soap init(&soap);
soap.proxy host = ”proxyhostname”;
soap.proxy port = 8080;
if (soap call ns method(&soap, ”http://host:port/path”, ”action”, ...))
soap print fault(&soap, stderr);
else
...
The attributes soap.proxy host and soap.proxy port keep their values through the remote method calls,
so they only need to be set once.
18.29
FastCGI Support
To enable FastCGI support, install FastCGI and compile all sources (including your application
sources that use stdsoap2.h) with option -DWITH FASTCGI or add
#define WITH FASTCGI
to stdsoap2.h.
18.30
How to Create gSOAP Applications With a Small Memory Footprint
To compile gSOAP applications intended for small memory devices, you may want to remove all
non-essential features that consume precious code and data space. To do this, compile the gSOAP
sources with -DWITH LEAN (i.e. #define WITH LEAN) to remove many non-essential features. The
features that will be disabled are:
• No I/O timeouts. Note that many socket operations already obey some form of timeout
handling, such as a connect timeout for example.
• No HTTP keep alive
• No HTTP cookies
• No HTTP authentication
• No HTTP chunked output (but input is OK)
223
• No HTTP compressed output (but input is OK when compiled with WITH GZIP)
• No send/recv timeouts
• No socket flags (no soap.socket flag, soap.connect flag, soap.bind flag, soap.accept flag)
• No canonical XML output
• No logging
• Limited TCP/IP and HTTP error diagnostic messages
• No support for time t serialization
• No support for LONG64/ULONG64 serialization (use typedef long xsd long)
Use -DWITH LEANER to make the executable even smaller by removing DIME and MIME attachment handling, wchar t* serialization, and support for XML DOM operations. Note that
DIME/MIME attachments are not essential to achieve SOAP/XML interoperability. DIME attachments are a convenient way to exchange non-text-based (i.e. binary) content, but are not
required for basic SOAP/XML interoperability. Attachment requirements are predictable. That is,
applications won’t suddenly decide to use DIME or MIME instead of XML to exchange content.
It is safe to try to compile your application with -DWITH LEAN, provided that your application does
not rely on I/O timeouts. When no linkage error occurs in the compilation process, it is safe to
assume that your application will run just fine.
18.31
How to Eliminate BSD Socket Library Linkage
The stdsoap2.c and stdsoap2.cpp gSOAP runtime libraries should be linked with a BSD socket library
in the project build, e.g. winsock for Win32. To eliminate the need to link a socket library, you can
compile stdsoap2.c (for C) and stdsoap2.cpp (for C++) with the -DWITH NOIO macro set (i.e. #define
WITH NOIO). This eliminates the dependency on the BSD socket API, IO streams, FILE type, and
errno.
As a consequence, you MUST define callbacks to replace the missing socket stack. To do so, add
to your code the following definitions:
struct soap soap;
soap init(&soap);
/* fsend is used to transmit data in blocks */
soap.fsend = my send;
/* frecv is used to receive data in blocks */
soap.frecv = my recv;
/* fopen is used to connect */
soap.fopen = my tcp connect;
/* fclose is used to disconnect */
soap.fclose = my tcp disconnect;
/* fclosesocket is used only to close the master socket in a server upon soap done() */
soap.fclosesocket = my tcp closesocket;
/* fshutdownsocket is used after completing a send operation to send TCP FIN */
224
soap.fshutdownsocket = my tcp shutdownsocket;
/* setting fpoll is optional, leave it NULL to omit polling the server */
soap.fpoll = my poll;
/* faccept is used only by a server application */
soap.faccept = my accept;
These functions are supposed to provide a (minimal) transport stack. See Section 18.7 for more
details on the use of these callbacks. All callback function pointers should be non-NULL, except
fpoll.
You cannot use soap print fault and soap print fault location to print error diagnostics. Instead, the
value of soap.error, which contains the gSOAP error code, can be used to determine the cause of a
fault.
18.32
How to Combine Multiple Client and Server Implementations into one
Executable
The wsdl2h tool can be used to import multiple WSDLs and schemas at once. The service definitions
are combined in one header file to be parsed by soapcpp2. It is important to assign namespace prefixes
to namespace URIs using the typemap.dat file. Otherwise, wsdl2h will assign namespace prefixes ns1,
ns2, and so on to the service operations and schema types. Thus, any change to a WSDL or schema
may result in a new prefix assignment. For more details, please see Section 7.2.11.
For example, consider the XMethods delayed stock quote and exchange rate services. We can
import both WSDLs at once with:
wsdl2h -s -o qx.h http://services.xmethods.net/soap/urn:xmethods-delayed-quotes.wsdlhttp://www.xmethods.net/sd/20
This generates the qx.h file for C++ (option -s for non-STL). The next step is to use the soapcpp2
compiler on this file to generate client code (option -C):
soapcpp2 -C qx.h
The resuling header file can be viewed with Doxygen to pretty-print the service definitions in
HTML.
You will notice that the qx.h file contains a very lengthy service name ”net x002exmethods x002eservices x002est
Since this is undesirable for naming files, operations, and proxies, we manually change its name
into ”StockQuote” and re-run soapcpp2.
This gives us a couple of files we need to build the application in C++:
soapStub.h
soapH.h
soapC.cpp
soapClient.cpp
soapStockQuoteProxy.h
soapCurrencyExchangeBindingProxy.h
StockQuote.nsmap
225
The two .cpp files and stdsoap2.cpp are compiled and linked with the main application:
#include ”soapStockQuoteProxy.h”
#include ”soapCurrencyExchangeBindingProxy.h”
#include ”StockQuote.nsmap”
main()
{
StockQuote stock;
CurrencyExchangeBinding exchange;
float quote, rate;
stock.ns1 getQuote(”IBM”, quote); // get quote for IBM
if (stock.soap->error)
exit(1);
exchange.ns2 getRate(”us”, ”uk”, rate); // get US to UK rate
if (exchange.soap->error)
exit(1);
cout << ”IBM in UK pounds = ” << rate*quote << endl;
}
This application prints the value of the IBM stock in UK pounds.
Note that the prefixes ns1 and ns2 can be changed using the typemap.dat file for wsdl2h, see Section 7.2.11.
Another approach to combine multiple client and service applications into one executable is by using
C++ namespaces to logically separate the definitions or by creating C libraries for the client/server
objects as explained in subsequent sections.
18.33
How to Build a Client or Server in a C++ Code Namespace
You can use a C++ code namespace of your choice in your header file to build a client or server in
that code namespace. In this way, you can create multiple clients and servers that can be combined
and linked together without conflicts, which is explained in more detail in the next section (which
also shows an example combining two client libraries defined in two C++ code namespaces).
At most one namespace can be defined for the entire gSOAP header file. The code namespace
MUST completely encapsulate the entire contents of the header file:
namespace myNamespaceName {
... gSOAP header file contents ...
}
When compiling this header file with the gSOAP compiler, all type definitions, the (de)serializers for
these types, and the stub/skeleton codes will be placed in this namespace. The XML namespace
mapping table (saved in a .nsmap file) will not be placed in the code namespace to allow it to
be linked as a global object. You can use option -n to create local XML namespace tables, see
Section 8.1 (but remember that you explicitly need to initialize the soap.namespaces to point to a
table at run time). The generated files are prefixed with the code namespace name instead of the
usual soap file name prefix to enable multiple client/server codes to be build in the same project
directory (a code namespace automatically sets the -p compiler option, see Section 8.1 for options).
226
Because the SOAP Header and Fault serialization codes will also be placed in the namespace, they
cannot be called from the stdsoap2.cpp run time library code and are therefore rendered unusable.
Therefore, these serializers are not compiled at all (enforced with #define WITH NOGLOBAL). To add
SOAP Header and Fault serializers, you MUST compile them separately as follows. First, create a
new header file env.h with the SOAP Header and Fault definitions. You can leave this header file
empty if you want to use the default SOAP Header and Fault. Then compile this header file with:
soapcpp2 -penv env.h
The generated envC.cpp file holds the SOAP Header and Fault serializers and you can link this file
with your client or server application.
18.34
How to Create Client/Server Libraries
The gSOAP compiler produces soapClientLib.cpp and soapServerLib.cpp codes that are specifically
intended for building static or dynamic client/server libraries. These codes export the stubs and
skeletons, but keep all marshaling code (i.e. parameter serializers and deserializers) local (i.e. as
static functions) to avoid link symbol conflicts when combining multiple clients and/or servers into
one executable. Note that it is far simpler to use the wsdl2h tool on multiple WSDL files to generate
a header file that combines all service definitions. However, the approach presented in this section
is useful when creating (dynamic) libraries for client and server objects, such as DLLs as described
in Section 18.35.
To build multiple libraries in the same project directory, you can define a C++ code namespace in
your header file (see Section 18.33) or you can use soapcpp2 with option -p to rename the generated
soapClientLib.cpp and soapServerLib.cpp (and associated) files. The -p option specifies the file name
prefix to replace the soap prefix. The libraries don’t have to be C++ codes. You can use option -c to
generate C code. A clean separation of libraries can also be achieved with C++ code namespaces,
see Section 18.33.
The library codes do not define SOAP Header and Fault serializers. You MUST add SOAP Header
and Fault serializers to your application, which are compiled separately as follows. First, create a
new header file env.h with the SOAP Header and Fault definitions. You can leave this header file
empty if you want to use the default SOAP Header and Fault. Then compile this header file with:
soapcpp2 -c -penv env.h
The generated envC.cpp file holds the SOAP Header and Fault serializers and you can create a
(dynamic) library for it to link this code with your client or server application.
You MUST compile the stdsoap2.cpp library using -DWITH NONAMESPACES:
g++ -DWITH NONAMESPACES -c stdsoap2.cpp
This omits the reference to the global namespaces table, which is nowhere to be defined since we
will use XML namespaces for each client/service separately. Therefore, you MUST explicitly set
the namespaces value of the gSOAP environment in your code every time after initialization of the
soap struct.
227
For example, suppose we have two clients defined in header files client1.h and client2.h. We first
generate the envH.h file for the SOAP Header and Fault definitions:
soapcpp2 -c -penv env.h
Then we generate the code for client1 and client2:
soapcpp2 -c -n -pmyClient1 client1.h
soapcpp2 -c -n -pmyClient2 client2.h
This generates myClient1ClientLib.c and myClient2ClientLib.c (among many other files). These two files
should be compiled and linked with your application. The source code of your application should
include the generated envH.h, myClient1H.h, myClient2.h files and myClient1.nsmap, myClient2.nsmap files:
#include ”envH.h” // include this file first!
#include ”myClient1H.h” // include client 1 stubs
#include ”myClient2H.h” // include client 2 stubs
...
#include ”myClient1H.nsmap” // include client 1 nsmap
#include ”myClient2H.nsmap” // include client 2 nsmap
...
soap init(&soap);
soap set namespaces(&soap, myClient1 namespaces);
... make Client 1 invocations ...
...
soap set namespaces(&soap, myClient2 namespaces);
... make Client 2 invocations ...
It is important to use soapcpp2 option -n, see Section 8.1, to rename the namespace tables so we
can include them all without running into redefinitions.
Note: Link conflicts may still occur in the unlikely situation that identical remote method names
are defined in two or more client stubs or server skeletons when these methods share the same
XML namespace prefix. You may have to use C++ code namespaces to avoid these link conflicts
or rename the namespace prefixes used by the remote method defined in the header files.
18.34.1
C++ Example
As an example we will build a Delayed Stock Quote client library and a Currency Exchange Rate
client library.
First, we create an empty header file env.h (which may contain optional SOAP Header and Fault
definitions), and compile it as follows:
soapcpp2 -penv env.h
g++ -c envC.cpp
We also compile stdsoap2.cpp without namespaces:
228
g++ -c -DWITH NONAMESPACES stdsoap2.cpp
Note: when you forget to use -DWITH NONAMESPACES you will get an unresolved link error for the
global namespaces table. You can define a dummy table to avoid having to recompile stdsoap2.cpp.
Second, we create the Delayed Stock Quote header file specification, which may be obtained using
the WSDL importer. If you want to use C++ namespaces then you need to manually add the
namespace declaration to the generated header file:
namespace quote {
//gsoap ns service name: Service
//gsoap ns service style: rpc
//gsoap ns service encoding: encoded
//gsoap ns service location: http://services.xmethods.net/soap
//gsoap ns schema namespace: urn:xmethods-delayed-quotes
//gsoap ns service method-action: getQuote ””
int ns getQuote(char *symbol, float &Result);
}
We then compile it as a library and we use option -n to rename the namespace table to avoid link
conflicts later:
soapcpp2 -n quote.h
g++ -c quoteClientLib.cpp
If you don’t want to use a C++ code namespace, you should compile quote.h “as is” with soapcpp2
option -pquote:
soapcpp2 -n -pquote quote.h
g++ -c quoteClientLib.cpp
Third, we create the Currency Exchange Rate header file specification:
namespace rate {
//gsoap ns service name: Service
//gsoap ns service style: rpc
//gsoap ns service encoding: encoded
//gsoap ns service location: http://services.xmethods.net/soap
//gsoap ns schema namespace: urn:xmethods-CurrencyExchange
//gsoap ns service method-action: getRate ””
int ns getRate(char *country1, char *country2, float &Result);
}
Similar to the Quote example above, we compile it as a library and we use option -n to rename the
namespace table to avoid link conflicts:
soapcpp2 -n rate.h
229
Fourth, we consider linking the libraries to the main program. The main program can import the
quoteServiceProxy.h and rateServiceProxy.h files to obtain client proxies to invoke the services. The
proxy implementations are defined in quoteClient.cpp. The -n option also affects the generation
of the C++ proxy codes to ensure that the gSOAP environment is properly initialized with the
appropriate namespace table (so you don’t have to initialize explicitly – this feature is only available
with C++ proxy and server object classes).
#include ”quoteServiceProxy.h” // get quote Service proxy
#include ”rateServiceProxy.h” // get rate Service proxy
#include ”quote.nsmap” // get quote namespace bindings
#include ”rate.nsmap” // get rate namespace bindings
int main(int argc, char *argv[])
{
if (argc <= 1)
{
std::cerr << ”Usage: main ticker [currency]” << std::endl
exit(0);
}
quote::Service quote;
float q;
if (quote.getQuote(argv[1], q)) // get quote
soap print fault(quote.soap, stderr);
else
{
if (argc > 2)
{
rate::Service rate;
float r;
if (rate.getRate(”us”, argv[2], r)) // get rate in US dollars
soap print fault(rate.soap, stderr);
else
q *= r; // convert the quote
}
std::cout << argv[1] << ”: ” << q << std::endl;
}
return 0;
}
Compile and link this application with stdsoap2.o, envC.o, quoteServerProxy.o, and rateServerProxy.o.
To compile and link a server object is very similar. For example, assume that we need to implement
a calculator service and we want to create a library for it.
namespace calc {
//gsoap ns service name: Service
//gsoap ns service style: rpc
//gsoap ns service encoding: encoded
//gsoap ns service location: http://www.cs.fsu.edu/ engelen/calc.cgi
//gsoap ns schema namespace: urn:calc
int ns add(double a, double b, double &result);
int ns sub(double a, double b, double &result);
230
int ns mul(double a, double b, double &result);
int ns div(double a, double b, double &result);
}
We compile this with:
soapcpp2 -n calc.h
The effect of the -n option is that it creates local namespace tables, and a modified calcServiceObject.h
server class definitions that properly initialize the gSOAP run time with the table.
#include ”calcServiceObject.h” // get Service object
#include ”calc.nsmap” // get calc namespace bindings
...
calc::Service calc;
calc.serve(); // calls request dispatcher to invoke one of the functions below
...
int calc::Service::add(double a, double b, double &result);
{ result = a + b; returnSOAP OK; }
int calc::Service::sub(double a, double b, double &result);
{ result = a - b; returnSOAP OK; }
int calc::Service::mul(double a, double b, double &result);
{ result = a * b; returnSOAP OK; }
int calc::Service::div(double a, double b, double &result);
{ result = a / b; returnSOAP OK; }
In fact, the calc::Service class is derived from the struct soap. So the environment is available as this,
which can be passed to all gSOAP functions that require a soap struct context.
18.34.2
C Example
This is the same example as above, but the clients are build with pure C.
First, we create an empty header file (which may contain optional SOAP Header and Fault definitions), and compile it as follows:
soapcpp2 -c -penv env.h
gcc -c envC.c
We also compile stdsoap2.c without namespaces:
gcc -c -DWITH NONAMESPACES stdsoap2.c
Second, we create the Delayed Stock Quote header file specification, which may be obtained using
the WSDL importer.
231
//gsoap ns service name: Service
//gsoap ns service style: rpc
//gsoap ns service encoding: encoded
//gsoap ns service location: http://services.xmethods.net/soap
//gsoap ns schema namespace: urn:xmethods-delayed-quotes
//gsoap ns service method-action: getQuote ””
int ns getQuote(char *symbol, float *Result);
We compile it as a library and we use options -n and -p to rename the namespace table to avoid
link conflicts:
soapcpp2 -c -n -pquote quote.h
gcc -c quoteClientLib.c
Third, we create the Currency Exchange Rate header file specification:
//gsoap ns service name: Service
//gsoap ns service style: rpc
//gsoap ns service encoding: encoded
//gsoap ns service location: http://services.xmethods.net/soap
//gsoap ns schema namespace: urn:xmethods-CurrencyExchange
//gsoap ns service method-action: getRate ””
int ns getRate(char *country1, char *country2, float *Result);
We compile it as a library and we use options -n and -p to rename the namespace table to avoid
link conflicts:
soapcpp2 -c -n -prate rate.h
gcc -c rateClientLib.c
The main program is:
#include ”quoteStub.h” // get quote Service stub
#include ”rateStub.h” // get rate Service stub
#include ”quote.nsmap” // get quote namespace bindings
#include ”rate.nsmap” // get rate namespace bindings
int main(int argc, char *argv[])
{
if (argc <= 1)
{
fprintf(stderr, ”Usage: main ticker [currency]\n”);
exit(0);
}
struct soap soap;
float q;
soap init(&soap);
soap set namespaces(&soap, quote namespaces);
if (soap call ns getQuote(&soap, ”http://services.xmethods.net/soap”, ””, argv[1], &q)) // get
quote
soap print fault(&soap, stderr);
232
else
{
if (argc > 2)
{
soap set namespaces(&soap, rate namespaces);
float r;
if (soap call ns getRate(&soap, ”http://services.xmethods.net/soap”, ””, ”us”, argv[2],
&r)) // get rate in US dollars
soap print fault(&soap, stderr);
else
q *= r; // convert the quote
}
printf(”%s: %f \n”, argv[1], q);
}
return 0;
}
Compile and link this application with stdsoap2.o, envC.o, quoteClientLib.o, and rateClientLib.o.
To compile and link a server library is very similar. Assuming that the server is named “calc”
(as specified with options -n and -p), the application needs to include the calcStub.h file, link the
calcServerLib.o file, and call calc serve(&soap) function at run time.
18.35
18.35.1
How to Create DLLs
Create the Base stdsoap2.dll
First, create a new header file env.h with the SOAP Header and Fault definitions. You can leave
this header file empty if you want to use the default SOAP Header and Fault. Then compile this
header file with:
soapcpp2 -penv env.h
The generated envC.cpp file holds the SOAP Header and Fault serializers, which need to be part of
the base library functions.
The next step is to create stdsoap2.dll which consists of the file stdsoap2.cpp and envC.cpp. This DLL
contains all common functions needed for all other clients and servers based on gSOAP. Compile
envC.cpp and stdsoap2.cpp into stdsoap2.dll using the C++ compiler option -DWITH NONAMESPACES
and the MSVC Pre-Processor definitions SOAP FMAC1= declspec(dllexport) and SOAP FMAC3= declspec(dllexport)
(or you can compile with -DWITH SOAPDEFS H and put the macro definitions in soapdefs.h). This
exports all functions which are preceded by the macro SOAP FMAC1 in the soapcpp2.cpp source file
and macro SOAP FMAC3 in the envC.cpp source file.
18.35.2
Creating Client and Server DLLs
Compile the soapClientLib.cpp and soapServerLib.cpp sources as DLLs by using the MSVC Pre-Processor
definitions SOAP FMAC5= declspec(dllexport) and SOAP CMAC= declspec(dllexport), and by using the
C++ compiler option -DWITH NONAMESPACES. This DLL links to stdsoap2.dll.
233
To create multiple DLLs in the same project directory, you SHOULD use option -p to rename the
generated soapClientLib.cpp and soapServerLib.cpp (and associated) files. The -p option specifies the
file name prefix to replace the soap prefix. A clean separation of libraries can also be achieved with
C++ namespaces, see Section 18.33.
Unless you use the client proxy and server object classes (soapXProxy.h and soapXObject.h where X
is the name of the service), all client and server applications MUST explicitly set the namespaces
value of the gSOAP environment:
soap init(&soap);
soap set namespaces(&soap, namespaces);
where the namespaces[] table should be defined in the client/server source. These tables are generated
in the .nsmap files. You can rename the tables using option -n, see Section 8.1.
18.36
gSOAP Plug-ins
The gSOAP plug-in feature enables a convenient extension mechanism of gSOAP capabilities.
When the plug-in registers with gSOAP, it has full access to the run-time settings and the gSOAP
function callbacks. Upon registry, the plug-in’s local data is associated with the gSOAP run-time.
By overriding gSOAP’s function callbacks with the plug-in’s function callbacks, the plug-in can
extend gSOAP’s capabilities. The local plug-in data can be accessed through a lookup function,
usually invoked within a callback function to access the plug-in data. The registry and lookup
functions are:
int soap register plugin arg(struct soap *soap, int (*fcreate)(struct soap *soap, struct soap plugin
*p, void *arg), void *arg)
void* soap lookup plugin(struct soap*, const char*);
Other functions that deal with plug-ins are:
int soap copy(struct soap *soap);
void soap done(struct soap *soap);
The soap copy function returns a new dynamically allocated gSOAP environment that is a copy of
another, such that no data is shared between the copy and the original environment. The soap copy
function invokes the plug-in copy callbacks to copy the plug-ins’ local data. The soap copy function
returns a gSOAP error code or SOAP OK. The soap done function de-registers all plugin-ins, so this
function should be called to cleanly terminate a gSOAP run-time environment.
An example will be used to illustrate these functions. This example overrides the send and receive
callbacks to copy all messages that are sent and received to the terminal (stderr).
First, we write a header file plugin.h to define the local plug-in data structure(s) and we define a
global name to identify the plug-in:
#include ”stdsoap2.h”
#define PLUGIN ID ”PLUGIN-1.0” // some name to identify plugin
struct plugin data // local plugin data
234
{
int (*fsend)(struct soap*, const char*, size t); // to save and use send callback
size t (*frecv)(struct soap*, char*, size t); // to save and use recv callback
};
int plugin(struct soap *soap, struct soap plugin *plugin, void *arg);
Then, we write the plugin registry function and the callbacks:
#include ”plugin.h”
static const char plugin id[] = PLUGIN ID; // the plugin id
static int plugin init(struct soap *soap, struct plugin data *data);
static int plugin copy(struct soap *soap, struct soap plugin *dst, struct soap plugin *src);
static void plugin delete(struct soap *soap, struct soap plugin *p);
static int plugin send(struct soap *soap, const char *buf, size t len);
static size t plugin recv(struct soap *soap, char *buf, size t len);
// the registry function:
int plugin(struct soap *soap, struct soap plugin *p, void *arg)
{
p->id = plugin id;
p->data = (void*)malloc(sizeof(struct plugin data));
p->fcopy = plugin copy; /* optional: when set the plugin must copy its local data */
p->fdelete = plugin delete;
if (p->data)
if (plugin init(soap, (struct plugin data*)p->data))
{
free(p->data); // error: could not init
return SOAP EOM; // return error
}
return SOAP OK;
}
static int plugin init(struct soap *soap, struct plugin data *data)
{
data->fsend = soap->fsend; // save old recv callback
data->frecv = soap->frecv; // save old send callback
soap->fsend = plugin send; // replace send callback with new
soap->frecv = plugin recv; // replace recv callback with new
return SOAP OK;
}
// copy plugin data, called by soap copy() // This is important: we need a deep copy to avoid data
sharing by two run-time environments
static int plugin copy(struct soap *soap, struct soap plugin *dst, struct soap plugin *src)
{
if (!(dst->data = (struct plugin data*)malloc(sizeof(struct plugin data))))
return SOAP EOM;
*dst->data = *src->data;
return SOAP OK;
}
// plugin deletion, called by soap done()
static void plugin delete(struct soap *soap, struct soap plugin *p)
{ free(p->data); // free allocated plugin data
}
235
// the new send callback
static int plugin send(struct soap *soap, const char *buf, size t len)
{
struct plugin data *data = (struct plugin data*)soap lookup plugin(soap, plugin id); // fetch
plugin’s local data
fwrite(buf, len, 1, stderr); // write message to stderr
return data->fsend(soap, buf, len); // pass data on to old send callback
}
// the new receive callback
static size t plugin recv(struct soap *soap, char *buf, size t len)
{
struct plugin data *data = (struct plugin data*)soap lookup plugin(soap, plugin id); // fetch
plugin’s local data
size t res = data->frecv(soap, buf, len); // get data from old recv callback
fwrite(buf, res, 1, stderr);
return res;
}
The fdelete callback of struct soap plugin MUST be set to register the plugin. It is the responsibility
of the plug-in to handle registry (init), copy, and deletion of the plug-in data and callbacks.
A plugin is copied with the soap copy() call. This function copies a soap struct and the chain of
plugins. It is up to the plugin implementation to share the plugin data or not:
1. if the fcopy() callback is set by the plugin initialization, this callback will be called to allow
the plugin to copy its local data upon a soap copy() call. When soap done() is called on the
soap struct copy, the fdelete() callback is called for deallocation and cleanup of the local data.
2. if the fcopy() callback is not set, then the plugin data will be shared (i.e. the data pointer
points to the same address). The fdelete() callback will not be called upon a soap done() on a
copy of the soap struct. The fdelete() callback will be called for the original soap struct with
which the plugin was registered.
The example plug-in should be used as follows:
struct soap soap;
soap init(&soap);
soap register plugin(&soap, plugin);
...
soap done(&soap);
Note: soap register plugin(...) is an alias for soap register plugin arg(..., NULL). That is, it passes NULL
as an argument to plug-in’s registry callback.
A number of example plug-ins are included in the gSOAP package’s plugin directory. Some of these
plug-ins are discussed.
18.36.1
The Message Logging and Statistics Plug-in
The message logging and access statistics plug-in can be used to selectively log inbound and outbound messages to a file or stream. It also keeps access statistics to log the total number of bytes
236
sent and received.
To use the plug-in, compile and link your application with logging.c located in the plugin directory of
the package. To enable the plug-in in your code, register the plug-in and set the streams as follows:
#include ”logging.h”
...
if (soap register plugin(&soap, logging))
soap print fault(&soap, stderr); // failed to register
...
struct logging data *logdata;
logdata = (struct logging data*)soap lookup plugin(&soap, logging id);
if (!logdata)
... // if the plug-in registered OK, there is certainly data but can’t hurt to check
logdata->inbound = stdout; // log to stdout
logdata->outbound = stdout; // log to stdout
... process messages ...
logdata->inbound = NULL; // don’t log
logdata->outbound = NULL; // don’t log
... process messages ...
size t bytes in = logdata->stat recv;
size t bytes out = logdata->stat sent;
If you use soap copy to copy the soap struct with the plug-in, the plug-in’s data will be shared by the
copy. Therefore, the statistics are not 100% guaranteed to be accurate for multi-threaded services
since race conditions on the counters may occur. Mutex is not used to update the counters to avoid
introducing expensive synchronization points. If 100% server-side accuracy is required, add mutex
at the points indicated in the logging.c code.
18.36.2
The HTTP GET Plug-in
The HTTP GET plug-in allows your server to handle HTTP GET requests as well as SOAP-based
POST request. HTTP GET requests can also be handled with the fget callback, see Section 18.7.
However, the HTTP GET plug-in also keeps statistics on the number of successful POST and GET
exchanges and failed operations (HTTP faults, SOAP Faults, etc.). It also keeps hit histograms
accumulated for up to a year.
To use the plug-in, compile and link your application with httpget.c located in the plugin directory
of the package. To enable the plug-in in your code, register the plug-in with your HTTP GET
handler function as follows:
#include ”httpget.h”
...
if (soap register plugin arg(&soap, httpget, (void*)my http get handler))
soap print fault(&soap, stderr); // failed to register
...
struct http get data *httpgetdata;
httpgetdata = (struct http get data*)soap lookup plugin(&soap, http get id);
if (!httpgetdata)
... // if the plug-in registered OK, there is certainly data but can’t hurt to check
237
... process messages ...
size t get ok = httpgetdata->stat get;
size t post ok = httpgetdata->stat post;
size t errors = httpgetdata->stat fail;
...
time t now = time(NULL);
struct tm *T;
T = localtime(&now);
size t hitsthisminute = httpgetdata->min[T->tm min];
size t hitsthishour = httpgetdata->hour[T->tm hour];
size t hitstoday = httpgetdata->day[T->tm yday];
An HTTP GET handler can simply produce HTML content, or any other type of information:
int my http get handler(struct soap)
{
soap->http content = ”text/html”;
soap response(soap, SOAP FILE);
soap send(soap, ”¡html¿Hello¡/html¿”);
soap end send(soap);
return SOAP OK; // return SOAP OK or HTTP error code, e.g. 404
}
If you use soap copy to copy the soap struct with the plug-in, the plug-in’s data will be shared by the
copy. Therefore, the statistics are not 100% guaranteed to be accurate for multi-threaded services
since race conditions on the counters may occur. Mutex is not used to update the counters to avoid
introducing expensive synchronization points. If 100% server-side accuracy is required, add mutex
at the points indicated in the httpget.c code.
18.36.3
The HTTP MD5 Plug-in
The HTTP MD5 plug-in works in the background to automatically verify the content of messages
using MD5 checksums. With the plug-in, messages can be transferred over (trusted but) unreliable
connections. The plug-in can be used on the client side and server side.
To use the plug-in, compile and link your application with httpmd5.c and md5evp.c located in the
plugin directory of the package. The md5evp.c implementation uses the EVP interface to compute
MD5 checksums with OpenSSL (compiled with -DWITH OPENSSL).
To enable the plug-in in your code, register the plug-in as follows:
#include ”httpmd5.h”
...
if (soap register plugin(&soap, http md5))
soap print fault(&soap, stderr); // failed to register
Once registered, MD5 checksums are produced for all outbound messages. Inbound messages with
MD5 checksums in the HTTP header are automatically verified.
The plug-in requires you to set the SOAP IO STORE flag when sending SOAP with attachments:
238
#include ”httpmd5.h”
...
struct soap soap;
soap init1(&soap, SOAP IO STORE);
if (soap register plugin(&soap, http md5)
soap print fault(&soap, stderr); // failed to register
... now safe to send SOAP with attachments ...
Unfortunately, this eliminates streaming.
18.36.4
The HTTP Digest Authentication Plug-in
The HTTP digest authentication plug-in enables a more secure authentication scheme compared
to basic authentication. HTTP basic authentication sends unencrypted userids and passwords
over the net, while digest authentication does not exchange passwords but exchanges checksums
of passwords (and other data such as nonces to avoid replay attacks). For more details, please see
RFC 2617.
The HTTP digest authentication can be used next to the built-in basic authentication, or basic
authentication can be rejected to tighten security. The server must have a database with userid’s
and passwords (in plain text form). The client, when challenged by the server, checks the authentication realm provided by the server and sets the userid and passwords for digest authentication.
The client application can temporarily store the userid and password for a sequence of message exchanges with the server, which is faster than repeated authorization challenges and authentication
responses.
At the client side, the plug-in is registered and service invocations are checked for authorization
challenges (HTTP error code 401). When the server challenges the client, the client should set the
userid and password and retry the invocation. The client can determine the userid and password
based on the authentication realm part of the server’s challenge. The authentication information
can be temporarily saved for multiple invocations.
Client-side example:
#include ”httpda.h”
...
if soap register plugin(&soap, http da))
soap print fault(&soap, stderr); // failed to register
...
if (soap call ns method(&soap, ...) != SOAP OK)
{
if (soap.error == 401) // challenge: HTTP authentication required
{
if (!strcmp(soap.authrealm, authrealm)) // determine authentication realm
{
struct http da info info; // to store userid and passwd
http da save(&soap, &info, authrealm, userid, passwd); // set userid and passwd for this
realm
if (soap call ns method(&soap, ...) == SOAP OK) // retry
{ ...
soap end(&soap); // userid and passwd were deallocated
239
http da restore(&soap, &info); // restore userid and passwd
if (!soap call ns method(&soap, ...) == SOAP OK) // another call
...
http da release(&soap, &info); // remove userid and passwd
This code supports both basic and digest authentication.
The server can challenge a client using HTTP code 401. With the plug-in, HTTP digest authentication challenges are send. Without the plug-in, basic authentication challenges are send.
Each server method can implement authentication as desired and may enforce digest authentication
or may also accept basic authentication responses. To verify digest authentication responses, the
server should compute and compare the checksums using the plug-in’s http da verify post function
for HTTP POST requests (and http da verify get for HTTP GET requests with the HTTP GET
plugin) as follows:
#include ”httpda.h”
...
if (soap register plugin(&soap, http da))
soap print fault(&soap, stderr); // failed to register
...
soap serve(&soap);
...
int ns method(struct soap *soap, ...)
{
if (soap->userid && soap->passwd) // client used basic authentication { // may decide not
to handle, but if ok then go ahead and compare info:
if (!strcmp(soap->userid, userid) &&
!strcmp(soap->passwd, passwd))
{ ... handle request ...
return SOAP OK;
}
}
else if (soap->authrealm && soap->userid) // Digest authentication
{
passwd = ... // database lookup on userid and authrealm to find passwd
if (!strcmp(soap->authrealm, authrealm) && !strcmp(soap->userid, userid))
{
if (!http da verify post(soap, passwd))
{ ... handle request ...
return SOAP OK;
}
}
}
soap->authrealm = authrealm; // set realm for challenge
return 401; // Not authorized, challenge digest authentication
}
18.36.5
The WS-Addressing Plug-in
The WSA WS-Addressing plug-in and the source code are extensively documented in the doc
directory of the gSOAP package. Please refer to the documentation included in the package.
240
18.36.6
The WS-Security Plug-in
The WSSE WS-Security plug-in and the source code are extensively documented in the doc directory
of the gSOAP package. Please refer to the documentation included in the package.
241